Corrosion.. Fundamentals, Testing, and Protection (10ed., ASM, 2003)(ISBN 0871707055)(2597s) opt

Corrosion: Fundamentals, Testing, and Protection ASM INTERNATIONAL The Materials Information Company ® Publication I...

Author: Covino B.S. | Jr. Cramer S.D. (eds.)

183 downloads 1777 Views 74MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Corrosion: Fundamentals, Testing, and Protection ASM INTERNATIONAL

The Materials Information Company

®

Publication Information and Contributors Introduction Corrosion: Fundamentals, Testing, and Protection was published in 2003 as Volume 13A of the ASM Handbook. The Volume was prepared under the direction of the ASM Handbook Committee.

Volume Editors The Volume Editors were Stephen D. Cramer and Bernard S. Covino, Jr.

Authors and Contributors Thomas A. Adler Albany Research Center, U.S. Department of Energy M.K. Adler Flitton Idaho National Engineering and Environmental Laboratory Vinod S. Agarwala Naval Air Systems Command Tatyana N. Andryushchenko Intel Corporation Peggy J. Arps University of California, Irvine Denise Aylor Naval Surface Warfare Center Robert Baboian RB Corrosion Service Christopher C. Berndt State University of New York, Stony Brook Marita L. Berndt Brookhaven National Laboratory Bennett P. Boffardi Bennett P. Boffardi and Associates, Inc. Stuart Bond TWI Ltd. Alan Bray Systems and Materials Research Consultancy Michiel P.H. Brongers CC Technologies Laboratories, Inc. Craig L. Brooks Analytical Processes/Engineered Solutions, Inc. Rudolph G. Buchheit The Ohio State University Kenneth C. Cadien Intel Corporation Richard E. Chinn Albany Research Center, U.S. Department of Energy Sean G. Corcoran Virginia Tech University Bernard S. Covino, Jr. Albany Research Center, U.S. Department of Energy

Bruce D. Craig MetCorr Stephen D. Cramer Albany Research Center, U.S. Department of Energy Chester Dacres Dacco Sciences Inc. Marek Danielewski AGH University of Science and Technology (Krakow) Guy Davis Dacco Sciences Inc. Sheldon Dean Dean Corrosion Technology Stephen C. Dexter University of Delaware David Dreisinger University of British Columbia James C. Earthman University of California, Irvine Peter Elliott Corrosion & Materials Consultancy Inc. E. Escalante National Institute of Science and Technology (Retired) Allen D. Feller Intel Corporation Paul B. Fischer Intel Corporation Gerald Frankel The Ohio State University James Fritz Technical Marketing Resources Aleksander Gil University of Mining & Metallurgy (Krakow) William A. Glaeser Battelle Columbus Richard D. Granata Florida Atlantic University Zbigniew Grzesik The Ohio State University Harvey Hack Northrop Grumman Corp. Harry R. Hanson Bay Engineering Christopher Hahin Illinois Department of Transportation Robert H. Heidersbach Dr. Rust, Inc. Gordon R. Holcomb Albany Research Center, U.S. Department of Energy Kyle T. Honeycutt Analytical Processes/Engineered Solutions, Inc. Francois Huet Université Pierre et Marie Curie Anthony E. Hughes Commonwealth Scientific & Industrial Research Organisation

Iwao Iwasaki University of Minnesota Vijay K. Jain Indian Institute of Technology Barnie P. Jones Oregon Department of Transportation Russell Jones Battelle Pacific Northwest Laboratories Robert M. Kain Consultant Russell D. Kane InterCorr International Incorporated Farida Kasumzade Progress Casting Group, Inc. Robert G. Kelly University of Virginia Robert J. Klassen Royal Military College of Canada Gerhardus H. Koch CC Technologies Laboratories, Inc. David Kolman Los Alamos National Laboratory Lorrie Krebs Dacco Sciences Inc. Jerome Kruger Johns Hopkins University Kyei-Sing Kwong Albany Research Center, U.S. Department of Energy Tom Langill American Galvanizers Association Ralph W. Leonard GalvoInfo Center Brenda J. Little Naval Research laboratory Carl E. Locke, Jr. University of Kansas Florian Mansfeld University of Southern California Philippe Marcus Ecole Nationale Supérieure de Chimie de Paris Richard Martin BJ Unichem Chemical Services Steven A. Matthes Albany Research Center, U.S. Department of Energy Thomas B. Mills Analytical Processes/Engineered Solutions, Inc. Anne E. Miller Intel Corporation Ted Mooney Finishing.com. Inc. Kevin M. Moore Energetics, Inc. James Moran Alcoa Technical Center

Makoto Nishimura Oak (Nippon) Co., Ltd. Paul M. Natishan Naval Research Laboratory James Noel University of Western Ontario Ricardo P. Nogueira Université Pierre et Marie Curie Bernard Normand Université Pierre et Marie Curie Kevin Ogle Usinor Research Joe H. Payer Case Western Reserve University Ignacio Perez Navair Bopinder S. Phull Consultant Jimmy D. Poindexter BJ Unichem Chemical Services Scott A. Prost-Domasky Analytical Processes/Engineered Solutions, Inc. Elie Protopopoff Ecole Nationale Supérieure de Chimie de Paris Robert A. Rapp The Ohio State University Vilupanur A. Ravi California Polytechnic Institute James C. Rawers Albany Research Center, U.S. Department of Energy Raúl B. Rebak Lawrence Livermore National Laboratory Izumi N. Reed Wayne Reitz North Dakota State University Anne Robbins Pierre R. Roberge Royal Military College of Canada John R. Scully University of Virginia A.J. Sedriks Office of Naval Research E. Bud Senkowski KTA-Tator Inc. Sadiq Shah Western Illinois University Barbara Shaw Pennsylvania State University David Shifler Naval Surface Warfare Center David W. Shoesmith University of Western Ontario David C. Silverman Argentum Solutions, Inc.

Raymund Singleton Singleton Corporation Susan Smialowska The Ohio State University Jack Snodgrass Alcoa Technical Center Narasi Sridhar Southwest Research Institute Kurt H. Stern Naval Research Laboratory James Stott CAPCIS Ltd. Hideaki Takahashi Hokkaido University Hisasi Takenouti Université Pierre et Marie Curie Kenneth B. Tator KTA-Tator Inc. Neil G. Thompson CC Technologies Laboratories, Inc. Garth R. Tingey Jack Tinnea Tinnea Associates Peter F. Tortorelli Oak Ridge National Laboratory Joseph H. Tylczak Albany Research Center, U.S. Department of Energy Kunigahalli Vasanth Naval Surface Warfare Center Lucien Veleva CINVESTAB-IPN Y. Paul Virmani Federal Highway Administration Mark C. Williams National Energy Technology Laboratory, U.S. Department of Energy Charles F. Windisch Pacific Northwest National Laboratory Michael Wolpers Henkel KgaA Ian Wright Oak Ridge National Laboratories Lietai Yang Southwest Research Institute Te-Lin Yau Te-Lin Yau Consultancy Steven Y. Yu 3M Małgorzata Ziomek-Moroz Albany Research Center, U.S. Department of Energy

Reviewers Robert S. Alwitt Boundary Technologies, Inc.

David E. Alman Albany Research Center, U.S. Dept. of Energy S.V. Babu Clarkson University Sean Brossia Southwest Research Institute Monica M. Chauviere ExxonMobil Research and Engineering Lichun Leigh Chen Engineered Materials Solutions O.V. Chetty Indian Institute of Technology T.C. Chevrot TotalFinaElf Gustavo Cragnolino Southwestern Research Institute Jim Crum Special Metals Corporation Craig V. Darragh The Timken Company Larry DeLashmit Blair Rubber Jim Divine ChemMet, Ltd. Barry Dugan Zinc Corp. of America Henry E. Fairman Cincinnati Metallurgical Consultants Robert Frankenthal Benjamin Fultz Bechtel Corp. Martin Gagne Noranda, Inc. Edward Ghali Laval University Brian Gleeson Iowa State University Larry D. Hanke Materials Evaluation & Engineering Incorporated Jeffrey A. Hawk Albany Research Center, U.S. Dept. of Energy Krista Heidersbach ChevronTexaco Dennis D. Huffman The Timken Company Fred Ienna Shell Global Solutions Tom Jack Nova Research and Technology Center Dwight Janoff FMC Technologies Mark Jaworoski United Technologies Research Center Kent. L. Johnson

Engineering Systems Incorporated Joanne Jones-Meehan U.S. Naval Research Laboratory Dwaine L. Klarstrom Haynes International Inc. R. Komanduri Oklahoma State University Paul J. Kovach Stress Engineering Services Incorporated Virginia M. Lesser Oregon State University Donald R. Lesuer Lawrence Livermore National Laboratory George J. Licina Structural Integrity Assoc. Eugene L. Liening Dow Chemical Company McIntyre R. Louthan Savannah River Tech Center Kenneth C. Ludema University of Michigan Stan P. Lynch Aeronautical and Maritime Research Laboratory (Australia) Gregory Makar Westvaco William L. Mankins Metallurgical Services Incorporated Ron E. Marrelli Conoco Phillips George Matzkanin TRI/NTIAC Stephen Maxwell Commercial Microbiology Gerald H. Meier University of Pittsburgh Bert Moniz DuPont Company Neville R. Moody Sandia Corporation John J. Moore Colorado School of Mines Bill Mullins U.S. Army John N. Murray Murray's et al. Robert M. O'Brien University of Oregon Tom O'Keefe University of Missouri (Rolla) Sankara Papavinasam CANMET Antoine Pourbaix Cebelcor Srinivasan Raghavan

University of Arizona Srikanth K. Raghunathan Nanomat Incorporated Robert A. Rapp The Ohio State University Anthony P. Reynolds University of South Carolina Joseph L. Rose Pennsylvania State University Brian J. Saldanha DuPont Company John R. Scully University of Virginia Ken-ichi Shimizu Keio University John A. Shreifels George Mason University Robert Silberstein Northrop Grumman Integrated Systems Theresa C. Simpson Bethlehem Steel Corp. Ron Skabo CH2M Hill Karl P. Staudhammer Los Alamos National Laboratory Jean Stockard University of Oregon Glenn Stoner University of Virginia James Strathman Portland State University S.R. Taylor University of Virginia Herman Terryn Vrije Universiteit Brussel Wen-Ta Tsai National Cheng Kung University Vilayanur V. Viswanathan Pacific Northwest National laboratory J. von Fraunhofer University of Maryland Robert Woods Zaclon, Inc. John F. Young J.F. Young International Inc. Gregory Ke Zhang Teck Cominco Metals

Foreword ASM International is pleased to publish ASM Handbook, Volume 13A, Corrosion: Fundamentals, Testing, and Protection, the first book in a two-volume revision of the landmark 1987 Metals Handbook, 9th Edition volume on corrosion. ASM Handbook, Volume 13A has been completely revised and updated to address the needs of ASM International members and the global technical community for current and comprehensive information on

corrosion principles, evaluation techniques, and protection methods. Advances in material science and corrosion technologies since the 1987 Corrosion volume was published have lessened some of the costs and degradation caused by corrosion. However, the systems that society relies on have increased in complexity during this time, so corrosion can have more far-reaching effects. Corrosion remains a multibillion-dollar problem that confronts nearly every engineer in every industry. ASM International is indebted to the Co-Chairs and Editors of this Handbook, Stephen D. Cramer and Bernard S. Covino, Jr., who had the vision and the drive to undertake the huge effort to update and revise the 1987 Corrosion volume. ASM Handbook, Volume 13A is the first fruit of their efforts; they are also leading the project to complete ASM Handbook, Volume 13B, Corrosion: Materials, Environments, and Industries, scheduled to publish in 2005. The Editors have done an outstanding job in organizing the project, in recruiting renowned experts to oversee sections and to write or revise articles, and in reviewing every manuscript. We are pleased with their vision to recruit authors from Canada, Mexico, France, Germany, United Kingdom, Poland, Japan, India, and Australia, as well as from the United States. This diverse community of volunteers, sharing their knowledge and experience, make this Volume truly an international effort. We thank the authors and reviewers of the 1987 Corrosion volume, which at the time was the largest, most comprehensive volume on a single topic ever published by ASM. This new edition builds upon that groundbreaking project. Thanks also go to the ASM Handbook Committee for their oversight and involvement, and to the ASM editorial staff for their tireless efforts. We are especially grateful to the nearly 200 authors and reviewers listed in the next several pages. Their willingness to invest their time and effort and to share their knowledge and experience by writing, rewriting, and reviewing articles has made this Handbook an outstanding source of information. Donald R. Muzyka, President, ASM International Stanley C. Theobald, Managing Director, ASM International

Preface The direct cost of corrosion in the United States was estimated to be $276 billion annually for 1998, or 3.1% of the 1998 U.S. gross domestic product of $8.79 trillion Ref 1. Of the industry sectors analyzed, utilities and transportation experienced the largest costs. The largest investment in corrosion control and protection strategies was in protective organic coatings. Indirect costs of corrosion, including lost productivity and corrosion-related overhead and taxes, when averaged over industry sectors, were roughly equal to or greater than the direct costs. In some cases they were substantially greater. For example, indirect corrosion costs related to the U.S. bridge infrastructure were estimated to be more than 10 times the $8.3 billion direct cost from bridge corrosion damage. Additional information is available in the article “Direct Costs of Corrosion in the United States” in this Volume. ASM Handbook, Volume 13A, Corrosion: Fundamentals, Testing, and Protection, is the first volume in a twovolume update, revision, and expansion of Corrosion, Volume 13 of the ninth edition Metals Handbook, published in 1987. The second volume—ASM Handbook, Volume 13B, Corrosion: Materials, Environments, and Industries—is to be published in 2005. The purpose of these two volumes is to represent the current state of knowledge in the field of corrosion and to provide a perspective on future trends in the field. Metals remain the major focus of the Handbook, but nonmetallic materials occupy a more prominent position that reflects their wide and effective use to solve problems of corrosion. Wet or aqueous corrosion remains the major focus, but dry or gaseous corrosion is discussed more fully, reflecting the increased importance of processes at elevated and high temperatures. ASM Handbook, Volume 13A recognizes the global nature of corrosion research and practice and the international level of corrosion activities and interactions required to provide cost-effective, safe, and environmentally sound solutions to materials problems in chemically aggressive environments. Twenty percent of the articles in Volume 13A did not appear in the 1987 Handbook. Authors from more than ten countries have contributed to Volume 13A. The table of contents has been translated into Spanish, French, Russian, Japanese, and Chinese to make the Handbook accessible to a diverse audience. Extensive references provide a road map to the corrosion literature and are augmented by Selected References that are a source of additional information. Information technology has changed dramatically since 1987, and the most significant occurrence has been the development of the Internet as an information resource. In response, ASM International has made the contents of this Handbook and others in the ASM Handbook series available on the Web. This Handbook also provides a

list, current at the time of publication, of significant data sources and of major national, international, academic, and government corrosion organizations and institutions that are accessible on the Web. Corrosion is described by well-known laws of thermodynamics, kinetics, and electrochemistry. The many variables that influence the behavior of a material in its environment can lead to a wide and complex range of performance, from the benign to the catastrophic. Understanding and avoiding detrimental corrosion is an interdisciplinary effort requiring knowledge of chemistry, electrochemistry, materials, engineering, and structures. All applications of engineered materials pivot on the fulcrum between environmental degradation, of which corrosion is a major element, and service or service life, with cost determining the point of balance. Costs are determined not in the spare confines of a material and its environment but in a complex landscape defined by technical, economic, social, environmental, health, safety, legal, and consumer constraints. This is illustrated by the experience of a Portland, OR Water Bureau engineer working to make way for a new light rail line along city streets Ref 2: …Construction conflicts are anticipated…, but day-to-day construction also alters the original design and corrosion control scheme of existing installations. As development occurs and utilities weave and cross, coatings are damaged, pipes are shorted, wires are cut, and test stations always seem to disappear…Work had to be sequenced and paced to minimize traffic interference… Environmental regulators were classifying the pavement as an engineered cap for brownfield and other contaminated areas…Utilities responded by characterizing the roadway as a constantly opening and closing zipper because we continually construct there… Corrosion control methods for urban areas must be designed for installation and operation in a congested environment that is constantly changing. This Handbook is organized into six major sections addressing corrosion fundamentals, testing, and protection. The first Section, “Fundamentals of Corrosion,” covers the theory of aqueous and gaseous corrosion from the thermodynamic and kinetic perspectives. It presents the principles of electrochemistry, the mechanisms of corrosion processes, and the methods for measuring corrosion rates in aqueous, molten salt, liquid metals, and gaseous environments. It introduces geochemical modeling as a means for characterizing and understanding corrosion in complex environments. While corrosion is usually associated with the environmental degradation of a material, this Section also describes ways in which corrosion is used for constructive or beneficial purposes. The second Section, “Forms of Corrosion,” describes how to recognize the different types of corrosion and the forces that influence them. It addresses uniform corrosion, localized corrosion, metallurgically influenced corrosion, mechanically assisted corrosion, environmentally induced cracking, and microbiologically influenced corrosion. The Section introduces the complex processes of wear-corrosion interactions that accelerate material deterioration at rates greater than those resulting from wear processes or corrosion processes alone. The third Section, “Corrosion Testing and Evaluation,” describes the planning of corrosion tests, evaluation of test results, laboratory corrosion testing, simulated service testing, and in-service techniques for damage detection and monitoring. It concludes by describing standard methods and practices for evaluating the various forms of corrosion. The fourth Section, “Methods of Corrosion Protection,” begins by discussing as a baseline the corrosion resistance of bulk materials. The Section continues with methods of corrosion protection, including surface treatments and conversion coatings, ceramic, glass and oxide coatings, metal coatings, coatings and linings, electrochemical corrosion control methods, and corrosion inhibitors. The fifth Section, “Designing for Corrosion Control and Prevention,” continues the theme of the fourth Section from the perspective of materials selection and equipment design. Corrosion control is an economic process as well as a technical process, and this Section discusses corrosion economic calculations, predictive modeling for structure service life, and a review of corrosion costs in the United States. The sixth Section, “Tools for the Corrosionist,” covers topics that are complementary to corrosion fundamentals, testing, and protection. It is a new addition to the Handbook. The topics include conventions and definitions in corrosion and oxidation, applications of modern analytical instruments in corrosion, materials science, statistics, and information sources and databases. Other useful Handbook contents include the “Glossary of Terms,” containing definitions of corrosion, electrochemistry, and materials terms common to corrosion and defined in the literature of ISO, ASTM, and NACE International. The “Corrosion Rate Conversion” Section includes conversions in both nomograph and tabular form. The metric conversion guide features conversion factors for common units and includes SI

prefixes. Finally, “Abbreviations and Symbols” provides a key to common acronyms, abbreviations, and symbols. The six Sections in the Handbook are divided into several subsections. These subsections were organized and written under the leadership of the following individuals (listed in alphabetical order): Chairperson Subsection Title Vinod S. Agarwala In-Service Techniques for Damage Detection and Monitoring Rudolph G. Buchheit Surface Treatments and Conversion Coatings Bernard S. Covino, Jr. Laboratory Corrosion Testing Bruce D. Craig Environmentally Induced Cracking Stephen D. Cramer Simulated Service Testing Metal Coatings Corrosion Inhibitors Tools for the Corrosionist Marek Danielewski Fundamentals of Gaseous Corrosion Stephen C. Dexter Microbiologically Influenced Corrosion Peter Elliott Designing for Corrosion Control and Protection Gerald Frankel Metallurgically Influenced Corrosion William A. Glaeser Mechanically Assisted Degradation Russell D. Kane Uniform Corrosion Carl E. Locke, Jr. Electrochemical Corrosion Control Methods Philippe Marcus Fundamentals of Corrosion Thermodynamics Paul M. Natishan Corrosion Resistance of Bulk Materials Bopinder S. Phull Evaluating Forms of Corrosion Vilupanur A. Ravi Ceramic, Glass, and Oxide Coatings Pierre R. Roberge Planning Corrosion Tests and Evaluating Results John R. Scully Fundamentals of Aqueous Corrosion Kinetics Susan Smialowska Localized Corrosion Kenneth B. Tator Coatings and Linings Peter F. Tortorelli Fundamentals Applied to Specific Environments Ian Wright Mechanically Assisted Degradation Margaret Ziomek-Moroz Fundamentals of Corrosion for Constructive Purposes These talented and dedicated individuals generously devoted considerable time to the preparation of this Handbook. They were joined in this effort by more than 120 authors who contributed their expertise and creativity in a collaborative venture to write or revise the articles and by more than 200 reviewers and 5 translators. These volunteers built on the contributions of earlier Handbook authors and reviewers who provided the solid foundation on which the present Handbook rests. For articles revised from the previous edition, the contribution of these authors is acknowledged at the end of articles. This location in no way diminishes their contribution or our gratitude. Those responsible for the current revision are named after the title. The variation in the amount of revision is broad. The many completely new articles presented no challenge for attribution, but assigning fair credit for revised articles was more problematic. The choice of presenting authors' names without comment or with the qualifier “Revised by” is solely the responsibility of the ASM staff. We thank ASM International and the ASM staff for their skilled support and valued expertise in the production of this Handbook. In particular, we thank Charles Moosbrugger, Gayle Anton, and Scott Henry for their encouragement, tactful diplomacy, and many discussions, plus, we should add, their wistful forbearance as deadlines came and went. The Albany Research Center, U.S. Department of Energy, gave us support and flexibility in our assignments to participate in this project and we are most grateful. In particular, we thank our supervisors Jeffrey A. Hawk and Cynthia P. Doğan, who were most gracious and generous with their encouragement throughout the project.

We close with these thoughtful words from T.R.B. (Tom) Watson, president of NACE International, 1964–65, author of Why Metals Corrode, and corrosion leader. (Ref 3) Mighty ships upon the ocean, suffer from severe corrosion. Even those that stay at dockside, are rapidly becoming oxide. Alas, that piling in the sea is mostly Fe2O3. And where the ocean meets the shore, you'll find there's Fe3O4. 'Cause when the wind is salt and gusty, things are getting awfully rusty. We can measure it, we can test it, we can halt it or arrest it; We can scrape it and weigh it; we can coat it or spray it; We can examine and dissect it; we can cathodically protect it. We can pick it up and drop it, but heaven knows we'll never stop it. So here's to rust, no doubt about it; most of us would starve without it. That said, given the thermodynamic, kinetic, and economic principles at work in our world, corrosion will not stop. This Handbook helps show us how to live with it. Stephen D. Cramer Bernard S. Covino, Jr. U.S. Department of Energy, Albany Research Center

References 1. G. H. Koch, M. P. H. Brongers, N. G. Thompson, Y. P. Virmani, and J. H. Payer, Corrosion Cost and Preventive Strategies in the United States, FHWA-RD-01–156, Federal Highway Administration, U.S. Department of Transportation, Washington D.C., 773 pp., March 2002. 2. Stu Greenberger, Underground Water Utilities – Crowded and Complex, Mater. Perform., Vol. 41, No. 7, July 2002, p. 8. 3. “Rust's a Must,” by T.R.B. Watson, poem reprinted by permission of Jean Watson; also reprinted in The Boatowner's Guide to Corrosion, by Everett Collier, Ragged Mountain Press, Camden ME, 2001. One line of the poem was modified for the purposes of this publication.

Officers and Trustees of ASM International (2002–2003) Donald R. Muzyka President and Trustee Special Metals Corporation (retired) Robert C. Tucker, Jr. Vice President and Trustee The Tucker Group, LLC Gordon H. Geiger Immediate Past President and Trustee University of Arizona John W. Pridgeon Treasurer Allvac Stanley C. Theobald Secretary and Managing Director ASM International Trustees Reza Abbaschian University of Florida Kathleen B. Alexander Los Alamos National Laboratory Rodney R. Boyer

Boeing Commercial Airplane Group Subi Dinda DaimlerChrysler Corporation R.G. (Gil) Gilliland Oak Ridge National Laboratory Andrew R. Nicoll Sulzer Metco (US) Inc. Richard D. Sisson, Jr. Worcester Polytechnic Institute George F. Vander Voort Buehler Ltd. Lawrence C. Wagner Texas Instruments Inc.

Members of the ASM Handbook Committee (2002–2003) Henry E. Fairman (Chair 2002–; Member 1993–) Cincinnati Metallurgical Consultants Jeffrey A. Hawk (Vice Chair 2002–; Member 1997–) U.S. Department of Energy David E. Alman (2002–) U.S. Department of Energy Bruce P. Bardes (1993–) Cincinnati Metallurgical Consultants Lichun Leigh Chen (2002–) Engineered Materials Solutions Craig V. Darragh (1989–) The Timken Company Larry D. Hanke (1994–) Materials Evaluation and Engineering Inc. Dennis D. Huffman (1982–) The Timken Company (retired) Dwight Janoff (1995–) FMC Corporation Kent L. Johnson (1999–) Engineering Systems Inc. Paul J. Kovach (1995–) Stress Engineering Services Inc. Donald R. Lesuer (1999–) Lawrence Livermore National Laboratory Huimin Liu (1999–) Ford Motor Company William L. Mankins (1989–) Metallurgical Services Inc. Srikanth Raghunathan (1999–) Nanomat Inc. Karl P. Staudhammer (1997–) Los Alamos National Laboratory Kenneth B. Tator (1991–) KTA-Tator Inc. George F. Vander Voort (1997–) Buehler Ltd.

George A. Wildridge (2000–) Borg Warner Morse TEC Corporation

Previous Chairs of the ASM Handbook Committee R.J. Austin (1992–1994) (Member 1984–) L.B. Case (1931–1933) (Member 1927–1933) T.D. Cooper (1984–1986) (Member 1981–1986) C.V. Darragh (1999–2002) (Member 1989–) E.O. Dixon (1952–1954) (Member 1947–1955) R.L. Dowdell (1938–1939) (Member 1935–1939) M.M. Gauthier (1997–1998) (Member 1990–) J.P. Gill (1937) (Member 1934–1937) J.D. Graham (1966–1968) (Member 1961–1970) J.F. Harper (1923–1926) (Member 1923–1926) C.H. Herty, Jr. (1934–1936) (Member 1930–1936) D.D. Huffman (1986–1990) (Member 1982–) J.B. Johnson (1948–1951) (Member 1944–1951) L.J. Korb (1983) (Member 1978–1983) R.W.E. Leiter (1962–1963) (Member 1955–1958, 1960–1964) G.V. Luerssen (1943–1947) (Member 1942–1947) G.N. Maniar (1979–1980) (Member 1974–1980) W.L. Mankins (1994–1997) (Member 1989–) J.L. McCall (1982) (Member 1977–1982) W.J. Merten (1927–1930) (Member 1923–1933) D.L. Olson (1990–1992) (Member 1982–1988, 1989–1992) N.E. Promisel (1955–1961) (Member 1954–1963) G.J. Shubat (1973–1975) (Member 1966–1975) W.A. Stadtler (1969–1972) (Member 1962–1972) R. Ward

(1976–1978) (Member 1972–1978) M.G.H. Wells (1981) (Member 1976–1981) D.J. Wright (1964–1965) (Member 1959–1967)

Staff ASM International staff who contributed to the development of the Volume included Charles Moosbrugger, Project Editor; Bonnie R. Sanders, Manager of Production; Gayle J. Anton, Editorial Assitant; Nancy Hrivnak, Jill Kinson, and Carol Polakowski, Production Editors; and Kathryn Muldoon, Production Assistant. Editorial Assistance was provided by Elizabeth Marquard, Heather Lampman, Mary Jane Riddlebaugh, and Beverly Musgrove. The Volume was prepared under the direction of Scott D. Henry, Assistant Director of Technical Publications and William W. Scott, Jr., Director of Technical Publications.

Preparation of Online Volume ASM Handbook, Volume 13A, Corrosion: Fundatmentals, Testing, and Protection, was converted to electronic files in 2004. The conversion was based on the First printing (2003). No substantive changes were made to the content of the Volume, but some minor corrections and clarifications were made as needed. ASM International staff who oversaw the conversion of the Volume to electronic files were Sally Fahrenholz-Mann, Sue Hess, and Susan Cheek. The electronic version was prepared under the direction of Stanley Theobald, Managing Director.

Copyright Information Copyright © 2003 by ASM International® All rights reserved No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the written permission of the copyright owner. First printing, October 2003 This book is a collective effort involving hundreds of technical specialists. It brings together a wealth of information from worldwide sources to help scientists, engineers, and technicians solve current and long-range problems. Great care is taken in the compilation and production of this Volume, but it should be made clear that NO WARRANTIES, EXPRESS OR IMPLIED, INCLUDING, WITHOUT LIMITATION, WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, ARE GIVEN IN CONNECTION WITH THIS PUBLICATION. Although this information is believed to be accurate by ASM, ASM cannot guarantee that favorable results will be obtained from the use of this publication alone. This publication is intended for use by persons having technical skill, at their sole discretion and risk. Since the conditions of product or material use are outside of ASM's control, ASM assumes no liability or obligation in connection with any use of this information. No claim of any kind, whether as to products or information in this publication, and whether or not based on negligence, shall be greater in amount than the purchase price of this product or publication in respect of which damages are claimed. THE REMEDY HEREBY PROVIDED SHALL BE THE EXCLUSIVE AND SOLE REMEDY OF BUYER, AND IN NO EVENT SHALL EITHER PARTY BE LIABLE FOR SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES WHETHER OR NOT CAUSED BY OR RESULTING FROM THE NEGLIGENCE OF SUCH PARTY. As with any material, evaluation of the material under end-use conditions prior to specification is essential. Therefore, specific testing under actual conditions is recommended. Nothing contained in this book shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in this book shall be construed as

a defense against any alleged infringement of letters patent, copyright, or trademark, or as a defense against liability for such infringement. Comments, criticisms, and suggestions are invited, and should be forwarded to ASM International. Library of Congress Cataloging-in-Publication Data ASM International ASM Handbook Includes bibliographical references and indexes Contents: v.1. Properties and selection—irons, steels, and high-performance alloys—v.2. Properties and selection—nonferrous alloys and special-purpose materials—[etc.]—v.21. Composites 1. Metals—Handbooks, manuals, etc. 2. Metal-work—Handbooks, manuals, etc. I. ASM International. Handbook Committee. II. Metals Handbook. TA459.M43 1990 620.1′6 90-115 SAN: 204-7586 ISBN: 0-87170-705-5 ASM International® Materials Park, OH 44073-0002 www.asminternational.org Printed in the United States of America Multiple copy reprints of individual articles are available from Technical Department, ASM International.

P. Marcus, Introduction to the Fundamentals of Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 3-4

Introduction to the Fundamentals of Corrosion Philippe Marcus, Ecole Nationale Supérieure de Chimie de Paris, Université Pierre et Marie Curie

Introduction THE SIGNIFICANT TECHNICAL CHALLENGES and the high cost directly related to corrosion provide strong incentives for engineers and other technical personnel to develop a firm grasp on the fundamental bases of corrosion. Understanding the fundamentals of corrosion is necessary not only for identifying corrosion mechanisms (a significant achievement by itself), but also for preventing corrosion by appropriate corrosion protection means and for predicting the corrosion behavior of metallic materials in service conditions. Understanding the mechanisms of corrosion is the key to the development of a knowledge-based design of corrosion resistant alloys and to the prediction of the long-term behavior of metallic materials in corrosive environments. Two major areas are usually distinguished in the corrosion of metals and alloys. The first area is where the metal or alloy is exposed to a liquid electrolyte, usually water, and thus typically called aqueous corrosion. The second area is where corrosion takes place in a gaseous environment, often called oxidation, high- temperature oxidation, or high-temperature corrosion, and called gaseous corrosion here. These two areas have been (and still are sometimes) referred to as wet corrosion and dry corrosion. This distinction finds its origin (and its justification) in some fundamental differences in the mechanisms, in particular the electrochemical nature of reactions occurring in aqueous solution (or in a nonaqueous electrolyte), as compared to the formation of thick oxide layers in air or other oxidizing atmospheres, at high temperature with fast transport processes by solidstate diffusion through a growing oxide. The separation between the two areas, however, should not be overemphasized, because there are also similarities and analogies, for example: • • •

The initial stages of reaction involve the adsorption of chemical species on the metal surface that can be described by the Gibbs equation for both liquid and gaseous environments. The nucleation and growth phenomena of oxide layers and other compounds The use of surface analytical techniques

The fundamental aspects of aqueous and gaseous corrosion are addressed in this first Section of the Handbook. Corrosion of metallic materials is generally detrimental and must be prevented, but if it is well understood and controlled, it can also be used in a powerful and constructive manner for electrochemical production of fine patterns on metal as well as on semiconductor surfaces. These constructive purposes also include electrochemical machining (down to the micro- or even the nanoscale), electrochemical and chemicalmechanical polishing, and anodes for batteries and fuel cells. These topics are also addressed in this Section.

P. Marcus, Introduction to the Fundamentals of Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 3-4 Introduction to the Fundamentals of Corrosion Philippe Marcus, Ecole Nationale Supérieure de Chimie de Paris, Université Pierre et Marie Curie

Thermodynamics

The thermodynamic aspects of corrosion, whether for aqueous corrosion or gaseous corrosion, are discussed first. This approach is logical because the object of thermodynamics is to examine the driving force for corrosion. Thermodynamics sets the framework of what is possible and what is not. It predicts the direction in which the changes of the system can occur. The only reactions that can take place spontaneously are those that will lower the energy of the system. If thermodynamic calculations predict that a reaction cannot occur, the reaction will indeed not occur. However, thermodynamics does not provide any information on the rate at which a reaction will occur; that is the area of kinetics. Corrosion in aqueous solution is an electrochemical process where the corroding metal is an electrode in contact with an electrolyte. The processes taking place at the metal surface are thus electrode processes, and such processes must be defined. There is also a need to define electrode potentials and the way they can be measured with reference electrodes. These points are covered in three articles in this Section: “Electrode Processes,” “Electrode Potentials,” and “Potential Measurements with Reference Electrodes.” With these basic notions defined, it is possible to explain the principle of potential-pH (E-pH) diagrams in the article “Potential versus pH (Pourbaix) Diagrams.” These diagrams allow us to visualize in a practical and easy way the domains of stability of chemical species (solid phases and dissolved species) and to know at a glance what corrosion reactions can occur in a given metal-solution system. In the article on potential-pH diagrams, the diagrams for high temperature water are also presented, as well as an extension of the concept to species adsorbed on metal surfaces. The more specific topic of thermodynamics of corrosion in molten salts is treated in a separate article, “Molten Salt Corrosion Thermodynamics.” The interesting contribution of geochemistry in modeling of stable chemical states in complex chemical environments is considered in the last article, “Geochemical Modeling,” in this group of articles on corrosion thermodynamics. It is not superfluous to reemphasize the fact that thermodynamics can only be a first step in the investigation of the corrosion behavior of a given metal/aqueous solution system. Thermodynamics provides the road map. Using it, the possible destinations are clearly known.


Kinetics Once thermodynamics are understood, it becomes necessary to undertake another task, namely to examine which reactions, among those reactions allowed by thermodynamics, will occur and at what rate. This subject is addressed by a group of articles on the fundamentals of aqueous corrosion kinetics. The first article, “Kinetics of Aqueous Corrosion,” examines the relations between the current and the potential associated with each of the two or more electrochemical reactions constituting the mixed system characteristic of corrosion. The current-potential (I-E) curves, the corrosion potential, and the corrosion current form the basis of the kinetic approach. Here activation energies of the reactions come into play. The concept of mixed potential is essential. Corrosion occurs if both an anodic and a cathodic reaction take place, each reaction involving chemical species that correspond to a different oxidation-reduction system. Because of this, the corrosion potential is not an equilibrium potential but rather is considered a mixed potential. On this ground, and without forgetting the framework set by thermodynamics, the mechanisms of corrosion in aqueous solution are discussed in the article “Aqueous Corrosion Reaction Mechanisms.” The latest advances in this area, achieved by an intense effort in research worldwide, are reviewed. Inevitably, areas in which the mechanisms are not yet fully elucidated still exist and justify present and future research. Passivation of metals and alloys, a phenomenon in which a thin protective layer of oxide or oxihydroxide is formed on the surface, is a major aspect of corrosion from the scientific as well as the engineering point of view. This phenomenon—discussed in the article “Passivity” in this Section—is also an area in which huge

progress has been made in recent years. The mechanisms of oxide film growth, the chemical composition and the chemical states, the crystallographic structure, and the semiconductor properties of passive films are now better understood. The acquired knowledge allows more accurate prediction of the long-term behavior of passive films and design of passive layers that are increasingly resistant to corrosion. It is an area where the contribution of surface chemical and structural analysis has been overwhelming. It is interesting to note that passivity is a good example of a case where potential-pH diagrams must be used with caution. Indeed, whereas the E-pH diagrams predict well the formation of passive oxides on copper, they predict the absence of passive film in acid solution on metals such as nickel, iron, and chromium, which are in fact well passivated. This is due to the very low dissolution rate of the oxides of these metals at low pH. Thanks to the formation of passive films rich in chromium oxide, passivation of alloys containing chromium, whether they are nickel-base or iron-base alloys (stainless steels), is a major feature of highly corrosion-resistant alloys. Finally, to define critical reaction paths, the appropriate experimental methods must be available to measure corrosion rates. These methods are presented in the last article, “Methods for Determining Aqueous Corrosion Reaction Rates,” in the group of articles on fundamentals of corrosion kinetics.


Gaseous Corrosion The next group of articles, on fundamentals of gaseous corrosion (often called oxidation in a broad sense, rather than the strict sense of the formation of an oxide), is organized in a similar manner to the group on aqueous corrosion. The group of articles starts with the “Introduction to the Fundamentals of Corrosion in Gases,” followed by a review of thermodynamics in the article “Thermodynamics of Gaseous Corrosion.” Here the environment is not conductive, and the ionic processes take place only on the metal surface and in or on the corrosion products. Thermodynamics defines and quantifies the driving force of the oxidation reaction; that is, the lowering of the energy level of the metal. The chemical equilibrium between gas phase (e.g., oxygen, water vapor, hydrogen sulfide) and compounds formed on the metal surface (oxide or sulfide in the examples cited earlier) can be represented by diagrams of the partial pressure of formation versus temperature. Such diagrams, called Ellingham diagrams, allow us to quickly visualize the domains of stability of different oxides. Similar diagrams have also been calculated for sulfides. The kinetics of gaseous corrosion of metals and alloys and the oxidation mechanisms are examined in two articles: “Kinetics of Gaseous Corrosion Processes” and “Gaseous Corrosion Mechanisms.” The different oxidation laws (i.e., the mass increase as a function of time) are presented in detail. The oxide layers developed at high temperature are quite thick (up to several micrometers) compared to the layers formed at room temperature because of kinetic limitations associated with diffusion in the solid state. The diffusion mechanisms, in which defects such as cation and anion vacancies play a key role, are discussed. The effect of impurities is also considered. The differences between the lattice constants of the substrate and the oxide lead to the existence of stresses at the interface and in the oxide. Such stresses play a major role in the properties of the oxide layers, and it is important to understand how they are generated. The role of dislocations is also an important aspect in high temperature oxidation. In some industrial applications, the corrosive environment is complex, and the corrosion products are not protective oxides. The phenomena related to corrosion by hot gases and combustion products are considered in the article “Gaseous Corrosion Mechanisms.” Appropriate experimental methods must be used to measure the kinetics of dry oxidation. Such methods are described in the last article, “Methods for Measuring Gaseous Corrosion Rates,” in this group of articles on the fundamentals of gaseous corrosion.

Corrosion in more specific environments— molten salts (the thermodynamics of which has been addressed in a previous article) and liquid metals—is considered in the next group of articles. Molten salts are addressed in two articles, “Corrosion by Molten Salts” and “Corrosion by Molten Nitrates, Nitrites, and Fluorides.” Under certain conditions, which are described in the article “Corrosion by Liquid Metals,” the liquid metal can penetrate into the grain boundaries of the metallic material at rates that are sometimes spectacular. A dangerous consequence is embrittlement of the metal. At this point, the reader has been provided with all the fundamental bases of thermodynamics and kinetics necessary to understand the mechanisms of corrosion of metals in aqueous solution and the mechanisms of oxidation at high temperature, including the relevant experimental methods.


Constructive Uses of Corrosion The fundamental electrochemical reactions of corrosion, in particular anodic dissolution, can be, if they are well understood and controlled, used in a very powerful way to design and fabricate patterns on metal surfaces. The article “Electrochemical Machining” describes the use of such reactions to produce surface architectures at scales down to the micrometer and now approaching the nanometer scale. The article “Electropolishing” reviews another use of corrosion for constructive purpose, where electrochemical reactions are used to produce finely polished metallic surfaces by controlled dissolution in an appropriate electrolyte. Chemical- mechanical polishing, often called chemicalmechanical planarization (CMP) is rapidly developing under the impetus of important applications in the microelectronic industry; these advances are described in the article “ChemicalMechanical Planarization for Semiconductors” in this Section. The combination of mechanical abrasion with dissolution of the surface allows the fast planarization of complex structures (e.g., with narrow copper interconnects). For this process to be efficient and industrially applicable, very good control of the attack taking place in the corrosive bath is necessary and requires a detailed understanding of the mechanisms of dissolution and passivation of the metal surfaces subjected to CMP. There are other areas of importance in classical and modern technologies that rely, from the mechanistic point of view, on controlled corrosion. These are described in the articles “Electrochemical Refining,” “Anodes for Batteries,” and “Fuel Cells,” which are the final articles in this Section of the Handbook.


Conclusions The fundamental aspects of corrosion, discussed in this Section, are indispensable, not only to understand the mechanisms, but also to control corrosion, to design appropriate means of corrosion protection, and to be able to predict the long-term corrosion behavior of metallic materials. The articles of this Section have been written

by leading scientists from the academic and industrial research world, and this Section represents a major review of the fundamentals of corrosion. It opens the way to the articles of the following Sections on the different forms of corrosion, including uniform corrosion, localized corrosion, and stress-corrosion cracking, on corrosion testing and evaluation, and finally on the methods of corrosion protection.

P. Marcus, Introduction to Fundamentals of Corrosion Thermodynamics, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 5

Introduction to Thermodynamics

Fundamentals

of

Corrosion

Philippe Marcus, Ecole Nationale Supérieure de Chimie de Paris, Université Pierre et Marie Curie

Introduction THE DRIVING FORCE of corrosion is the lowering of energy associated with the oxidation of a metal. Thermodynamics examines and quantifies this driving force. It predicts if reactions can or cannot occur (i.e., if the metal will corrode or be stable). It does not predict at what rate these changes can or will occur: this is the area of kinetics. However, knowing from thermodynamics what reactions are possible is a necessary step in the attempt to understand, predict, and control corrosion.

P. Marcus, Introduction to Fundamentals of Corrosion Thermodynamics, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 5 Introduction to Fundamentals of Corrosion Thermodynamics Philippe Marcus, Ecole Nationale Supérieure de Chimie de Paris, Université Pierre et Marie Curie

Electrochemical Reactions Corrosion of metals and alloys in aqueous environments or other ionically conducting liquids is almost always electrochemical in nature. It occurs when two or more electrochemical reactions take place on a metal surface. One of these reactions results in the change of the metal or some elements in the metal alloy from a metallic state into a nonmetallic state. The products of corrosion may be dissolved species or solid corrosion products. Because electrochemical reactions are at the origin of corrosion, the corroding metal surface is considered an electrode. The ionically conducting liquid is the electrolyte in which the reactions take place. Different aspects of these reactions are considered in the articles “Electrode Processes,” “Electrode Potentials,” and “Potential Measurements with Reference Electrodes” in this Section: •

•

The structure of the electrode/electrolyte interface: There is a separation of charges between electrons in the metal and ions in the electrolyte, creating an electrically charged double layer. The ions in the solution interact with water molecules. Adsorption of ions on the electrode surface may also occur. Transport of chemical species: This takes place through the double layer and in the electrolyte by diffusion.

•

•

The potential difference across the electrode/ electrolyte interface: Electrode potentials need to be measured to evaluate the corrosion behavior of a metal (this is true for both the thermodynamics and the kinetics of corrosion). Potential measurements require the use of reference electrodes. The processes governing corrosion: These are electrode processes, involving oxidation and reduction reactions (or anodic or cathodic reactions). The corroding system does not produce any net charge and, thus, the electrons produced by the electrochemical oxidation of the metal (the anodic reaction) must be consumed by an electrochemical reduction reaction (the cathodic reaction).

Potential-pH (E-pH) diagrams (Pourbaix diagrams), based on equilibrium thermodynamics for metal-water systems, show at a glance the regions of stability of the various phases that can exist in the system. Their principle and their construction are presented in the article “Potential versus pH (Pourbaix) Diagrams” for binary metal-water systems and for ternary metal-additive-water systems. Their applications as well as their limitations are also discussed. The E-pH diagrams are very useful as a thermodynamic framework for kinetic interpretation, but they do not provide information on corrosion rates. The prediction of the corrosion behavior of metals in aqueous solutions at high temperatures is also of considerable importance for different technologies (including power generation systems in general and nuclear power systems in particular). Potential pH diagrams for metals in high temperature water are presented. The principle of E-pH diagrams has been extended to the case of adsorbed species on metal surfaces in water. Whereas the solid compounds treated in the classical diagrams are three-dimensional (bulk) compounds (e.g., oxides, hydroxides, sulfides), the formation of more stable two-dimensional adsorbed phases has been considered only recently. Due to the large energy of adsorption, adsorbed layers may form under E- pH conditions in which the usual solid compounds are thermodynamically unstable, and these adsorbed layers can induce marked changes in the corrosion behavior of metals. The method of calculation of the equilibrium potentials of species adsorbed on an electrode surface in water is presented, together with examples of applications, in a section on E-pH diagrams for adsorbed species.


Molten Salt Corrosion Thermodynamics Molten salts constitute a special environment of importance in a number of technologies. There is no solvent analogous to water in molten salts. A specific potential scale must be established for each medium, as described in the article “Molten Salt Corrosion Thermodynamics” in this Section. Equilibrium diagrams, analogous to the E-pH diagrams for aqueous solutions, have been constructed. In these diagrams, the equilibrium potential, E, is plotted as a function of the activity of oxide in the melt, expressed as pO2- = -log O2-. The thermodynamics of molten salt electrochemical cells is treated here.


Geochemical Modeling The article “Geochemical Modeling” discusses modeling software that has been developed by geochemists to describe the chemical state of local environments, with interesting and important applications in corrosion. These include nuclear waste storage, atmospheric corrosion involving environmental effects on corrosion product stability, and corrosion in elevated and high-temperature aqueous systems.

Electrode Processes Revised by E. Protopopoff and P. Marcus, Laboratoire de Physico-Chimie des Surfaces, CNRS, Ecole Nationale Supérieure de Chimie de Paris, Université Pierre et Marie Curie

Introduction ELECTROCHEMICAL, OR ELECTRODE, REACTIONS are reactions that occur with charge transfer between neutral or ionic reactants and a conducting material, called the electrode, acting as an electron source or an electron sink (Ref 1). Electrochemical reactions involve change in valence; that is, oxidation or reduction of the reacting elements. Oxidation and reduction are commonly defined as follows. Oxidation is the removal of electrons from atoms or groups of atoms, resulting in an increase in valence, and reduction is the addition of electrons to atoms or groups of atoms, resulting in a decrease in valence (Ref 2).

References cited in this section 1. J.O'M. Bockris, A.K.N. Reddy, and M. Gamboa-Aldeco, Modern Electrochemistry, Kluwer Academic/Plenum Publishers, New York, 2000 2. L. Pauling, General Chemistry, W.H. Freeman, 1964, p 338–360


Electrode Reactions Because electrochemical reactions occur in an electrochemical cell with oxidation reactions occurring at one electrode and reduction reactions occurring at the other electrode, they are often further defined as either cathodic reactions or anodic reactions. By definition, cathodic reactions are those types of reactions that result in reduction, such as: M2+(aq) + 2e- → M(s)

(Eq 1)

Anodic reactions are those types of reactions that result in oxidation, such as: M(s) → M2+(aq) + 2e-

(Eq 2)

Because of the production of electrons during oxidation and the consumption of electrons during reduction, oxidation and reduction are coupled events. If the ability to store large amounts of electrons does not exist,

equivalent processes of oxidation and reduction will occur together during the course of normal electrochemical reactions. The oxidized species provide the electrons for the reduced species. Electrode Processes Revised by E. Protopopoff and P. Marcus, Laboratoire de Physico-Chimie des Surfaces, CNRS, Ecole Nationale Supérieure de Chimie de Paris, Université Pierre et Marie Curie

The Electric Field near the Electrode The examples stated earlier, like many aqueous corrosion situations, involve the reaction of aqueous metal species at a metal electrode surface. The metal-aqueous solution interface is complex, as is the mechanism by which the reactions take place across the interface. Because the reduction-oxidation reactions involve ionic species in the electrolyte reacting at or near the metal surface, the electrode surface is charged relative to the solution and the reactions are associated with specific electrode potentials. The charged interface results in an electric field that extends into the solution and has a dramatic effect. A solution that contains water as the primary solvent is affected by the electric field near the metal because of its structure. Water is polar and can be visualized as dipolar molecules that have a positive side (hydrogen atoms) and a negative side (oxygen atoms). In the electric field caused by the charged interface, the water molecules act as small dipoles and align themselves in the direction of the electric field. Electrode Processes Revised by E. Protopopoff and P. Marcus, Laboratoire de Physico-Chimie des Surfaces, CNRS, Ecole Nationale Supérieure de Chimie de Paris, Université Pierre et Marie Curie

Solvation of Ions Ions that are present in the solution are charged because of the loss or gain of electrons. The positive charged ions (cations) and negative charged ions (anions) also have an electric field associated with them. The solvent (water) molecules act as small dipoles; therefore, they are also attracted to the charged ions and align themselves in the electric field established by the charge of the ion. Because the electric field is strongest close to the ion, some water molecules reside very close to an ionic species in solution. The attraction is great enough that these water molecules travel with the ion as it moves through the solvent. The tightly bound water molecules are referred to as the primary water sheath of the ion. The electric field is weaker at distances outside the primary water sheath, but it still disturbs the polar water molecules as the ion passes through the solution. The water molecules that are disturbed as the ion passes, but do not move with the ion, are usually referred to as the secondary water sheath. Figure 1 shows a representation of the primary and secondary solvent molecules for a cation in water. Because of their smaller size relative to anions, cations have a stronger electric field close to the ion and more water molecules are associated in their primary water sheath. Anions have few, if any, primary water molecules. A detailed description of the hydration of ions in solution is given in Ref 1.

Fig. 1 Schematic of the primary and secondary solvent molecules around a cation in water Ions and polar water molecules are attracted to the metal-aqueous solution interface of an electrode because of the strong electric field in this region. Water molecules form a first row at the metal surface. This row of water molecules limits the distance to which hydrated ions can approach the interface. Figure 2 shows a schematic diagram of a charged interface and the locations of cations at the surface. Also, the primary water molecules associated with the ionic species limit the distance the cations can approach. The plane of positive charge containing the cations closest to a negatively charged surface is thus at a fixed distance from the metal. This plane of charge is referred to as the outer- Helmholtz plane (OHP).

Fig. 2 Schematic of a charged interface and of the locations of cations at the electrode surface. OHP, outer-Helmholtz plane The region of the interface with charge separation (Fig. 3) is called the electric double layer and can be represented as a charged capacitor (Ref 1). The potential drop across the interface is also often simplified as a linear change in potential from the metal surface to the OHP.

Fig. 3 The simplified electric double layer at a metal-aqueous solution interface and equivalent capacitor. OHP, outer-Helmholtz plane The electric equivalent of a metal-aqueous solution interface where no reactions with electron transfer occur over a large range of potentials is a simple capacitor (Fig. 3). The electrode is then ideally polarizable. This is the case of mercury in poorly reactive electrolytes (Ref 1). Noble metals like platinum or gold behave as ideal polarizable electrodes within a limited range of potential. However, on most metals, especially the corrodible ones, reactions with electron transfer across the interface (electrochemical reactions) occur, leading to charge transfer currents, or faradic current. The interface is a barrier to the transfer of electrons from or to the metal; this can be represented by a resistance called the chargetransfer resistance. This resistance is not a simple ohmic resistance, because it varies with the electrode potential. It is constant only in a limited potential range. The electrical circuit equivalent to a metal-electrolyte interface with charge transfer is thus a parallel combination of a double-layer capacitance (CDL) and a charge transfer resistance (RCT) (Fig. 4) (Ref 1). It must be noted that this is the electrical equivalent of a very simple interface; for example, the circuit equivalent to an electrode covered by an oxide film is more complicated (Ref 1).

Fig. 4 Electric circuit equivalent to a metal-electrolyte interface showing the double-layer capacitance (CDL) in parallel with the charge transfer resistance (RCT). RE is the ohmic resistance of the electrolyte (Ref 1). If there were no difficulty in the transfer of electrons across the interface, the only resistance to the electron flow would be the diffusion of aqueous species to and from the electrode. The electrode would then be ideally nonpolarizable, and its potential would not change until the solution was deficient in electron acceptors and/or donors. This property is sought after for reference electrodes (see the article in this Section, “Potential Measurements with Reference Electrodes”). However, when dealing with the kinetics of electrochemical reactions (see the article “Kinetics of Aqueous Corrosion” in this Volume), the metal-electrolyte interface represents an energy barrier that must be overcome. Thus, reactions at the interface are often dominated by activated processes, and control of activation by polarization plays a significant role in electrochemical kinetics. A key in controlling corrosion consists in minimizing the kinetics of the anodic reaction, that is, dissolution of the metal, or the kinetics of the cathodic reactions.

Reference cited in this section 1. J.O'M. Bockris, A.K.N. Reddy, and M. Gamboa-Aldeco, Modern Electrochemistry, Kluwer Academic/Plenum Publishers, New York, 2000


Acknowledgment This article has been adapted from the article by Charles A. Natalie, Electrode Processes, Corrosion, Vol 13, ASM Handbook (formerly 9th ed. Metals Handbook), ASM International, 1987, p 18–19. Electrode Processes Revised by E. Protopopoff and P. Marcus, Laboratoire de Physico-Chimie des Surfaces, CNRS, Ecole Nationale Supérieure de Chimie de Paris, Université Pierre et Marie Curie

References 1. J.O'M. Bockris, A.K.N. Reddy, and M. Gamboa-Aldeco, Modern Electrochemistry, Kluwer Academic/Plenum Publishers, New York, 2000 2. L. Pauling, General Chemistry, W.H. Freeman, 1964, p 338–360

E. Protopopoff and P. Marcus, Electrode Potentials, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 8–12

Electrode Potentials E. Protopopoff and P. Marcus, Ecole Nationale Supérieure de Chimie de Paris, Université Pierre et Marie Curie

Introduction ONE OF THE IMPORTANT FEATURES of the electrified interface between the electrode and the electrolyte in the aqueous corrosion of metals is the existence of a potential difference across the double layer, which leads to the definition of the electrode potential. The electrode potential is one of the most important parameters in both the thermodynamics and the kinetics of corrosion. The fundamentals of electrode potentials are discussed in this article. Examples of the calculations of the potential at equilibrium are given in the article “Potential versus pH (Pourbaix) Diagrams” in this Section of the Volume.

E. Protopopoff and P. Marcus, Electrode Potentials, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 8–12 Electrode Potentials E. Protopopoff and P. Marcus, Ecole Nationale Supérieure de Chimie de Paris, Université Pierre et Marie Curie

Thermodynamics of Chemical Equilibria The object of chemical thermodynamics is to develop a mathematical treatment of the chemical equilibria and the driving forces behind chemical reactions. The desire is to catalog quantitative data concerning known equilibria that later can be used to predict other equilibria. The driving force for chemical reactions at constant pressure is the Gibbs free energy change, which has been expressed in thermodynamic treatments as the balance between the effects of energy (enthalpy) change and entropy change. The entropy of a system is related to the number of ways in which microscopic particles can be distributed among states accessible to them. The enthalpy (H), the entropy (S), and the free energy (G) are thermodynamic state functions (Ref 1, 2). Free Energy and Chemical Potential. The Gibbs free energy (G) is the result of its enthalpy (H) and its entropic factor (TS): G = H - TS, where T is the absolute temperature (in Kelvin) (Ref 2). The partial molar free energy for a substance A is equal to , where nA and ni are the number of moles of A and any other constituent i and P is the pressure. It is usually called chemical potential (μA) and depends not only on the chemical formula of the species involved but also on its activity (aA) (corrected concentration—see end of this section): (Eq 1) where R is the gas constant (8.3143 J/K · mol), and is the standard chemical potential of A; that is, the chemical potential of A in an arbitrarily selected state. If A is a pure condensed substance, aA is equal to unity; if A is a species in solution, aA is equal to the product of the concentration of A (usually molality mA in number of moles of A per kilogram of solvent) by an activity coefficient (γA) representing the deviation from ideal behavior. The standard state for a gaseous substance is the substance under a pressure such that its fugacity is 1 bar (0.1 MPa), for most gases a pressure of ~1 atm. The standard state for a condensed substance is the pure substance at the given temperature (T) and under pressure (P). The standard state for a dissolved species is the is dependent on the hypothetical ideal solution of the species with unit molality (mol/kg) at T and P. The temperature and, for a condensed substance, also on the pressure (this dependence is small and usually neglected for pressures not too far from 1 bar). The usual convention is to assign the value 0 to the chemical potentials of the elements in their stable form and standard state at 25 °C (77 °F), for example, pure gaseous H 2 and O2 at fugacity equal to 1 bar, solvated protons at activity unity. Hence, the standard chemical potential of a substance is equal to its standard Gibbs free energy of formation from its elements at 25 °C (77 °F) under 1 bar pressure. Law of Chemical Equilibria. Consider the following chemical reaction: ΣνRR ↔ ΣνPP

(Eq 2)

where R and P designate the reactants and products, respectively, and νR and νP are the associated stoichiometric coefficients. The molar Gibbs free energy change for this reaction is: ΔrG = ΣνPμP - ΣνRμR

(Eq 3)

where μP and μR are the chemical potentials of the products and reactants, respectively. If the chemical potentials are expressed as in Eq 1, one obtains:

(Eq 4)

where R is the gas constant, Π is the symbol for multiplying a series of terms, aP and aR are the activities of the products and reactants, respectively, and ΔrG0 is the standard free energy change for the reaction: (Eq 5) When the reaction is in thermodynamic equilibrium (there is no tendency for the reaction to proceed either forward or backward), it has been shown that its molar Gibbs free energy change (ΔrG) is equal to zero (Ref 2). Then, from Eq. 4 the classic equilibrium law is obtained: (Eq 6) where Keq is the equilibrium constant for the reaction at the temperature T: (Eq 7) The standard chemical potentials, or free energies of formation, of an extensive number of compounds have been cataloged for various temperatures. Typical data sources are listed as the Selected References. This allows the prediction of standard free energy changes and equilibrium constants for reactions over a wide range of conditions.

References cited in this section 1. J.M. Smith and H.C. Van Hess, Introduction to Chemical Engineering Thermodynamics, McGraw-Hill, 1975, p 159–162 2. K. Denbigh, Principles of Chemical Equilibrium, 2nd ed., Cambridge Press, 1981, p 133–186


Reactions in Aqueous Solution Galvanic Cell Reactions. If a strip of metal is placed in water, some metal atoms will be oxidized into hydrated (solvated) ions. Because of the electrons remaining in the metal (called an electrode), the positively charged metallic ions will remain very close to the negatively charged metal strip in a double layer, where a very high potential gradient exists, as described in the article “Electrode Processes” in this Volume. Thus, a potential difference will exist between the metal and the solution. This potential difference is not measurable, but it is possible to measure the potential difference between this system and another metal-ion system. If a zinc electrode placed in a solution of zinc ions is linked through an electric circuit including a voltmeter to a copper electrode placed in a solution of copper ions (Fig. 1), a positive potential difference will be measured between the copper electrode and the zinc electrode. This is evidence that copper is a more noble metal than zinc and has

less tendency to oxidize. This can be checked directly by the chemical reaction in Eq 8; if zinc is put into a solution containing copper ions, zinc will dissolve, while copper will deposit from its ions on to the zinc. This metal-displacement reaction is due to the oxidation-reduction reaction between zinc metal and copper ions: Zn(s) + Cu2+(aq) → Zn2+(aq) + Cu(s)

(Eq 8)

Fig. 1 Typical electrochemical cell (a) used to study the free energy change that accompanies electrochemical or corrosion reactions. In this example, the cell contains copper and zinc electrodes in equilibrium, with their ions separated by a porous membrane to mitigate mixing. For purposes of simplicity, the concentration of metal ions is maintained at unit activity; that is, each solution contains approximately 1 mole of metal ion per liter. The reactions for copper and zinc electrodes in each half-cell are given in Eq 9 and 10. However, at equilibrium, metal dissolution and deposition occur at each electrode at equal rates (r1 = r2), as shown in (b), which illustrates copper atoms being oxidized to cupric ions and, at other areas, cupric ions being reduced to metallic copper. Source: Ref 3 This reaction takes place spontaneously, because the standard free energy change associated with it (ΔrG0 = + − − ) is negative. Zinc metal will react with copper ions almost to completion; the reaction will stop only when the concentration of copper ions is very small and such that equilibrium is reached, with ΔrG = ΔrG0 + RT ln = 0. If the reverse procedure is tried, that is, copper metal placed in a solution containing zinc ions, because it is associated with a strongly positive standard free energy change, the reaction will occur to only a very small extent, with the reaction stopping when a certain very small concentration of copper ions has been produced, such that equilibrium is reached. The same reaction may be studied in an electrochemical cell, such as the one shown and described in Fig. 1(a). An electric circuit links a copper electrode in a solution of copper sulfate to a zinc electrode in a solution of zinc sulfate. If the external conduction path is closed, the oxidation-reduction reaction in Eq 8 will take place but under the form of two spatially separated electron-transfer reactions occurring each at one electrode, which are: •

An oxidation-dissolution reaction involving the extraction of electrons from outer metal atoms to form metallic ions at the zinc electrode, called the anode: Zn(s) → Zn2+(aq) + 2e-

(Eq 9)

•

A reduction-deposition reaction involving the addition of electrons to the metallic ions to form metal atoms deposited at the copper electrode, called the cathode: Cu2+ (aq) + 2e- → Cu(s)

(Eq 10)

Electrons will flow from the zinc anode where they are produced to the copper cathode where they are consumed; the two half-reactions in Eq 9 and 10 naturally combine in the electrochemical cell to form the oxidation-reduction reaction of Eq 8, termed the chemical cell reaction. They are referred to as electrochemical, electrode, or half-cell reactions. The requirement for writing the equation of the overall reaction by combination of the oxidation and reduction half-cell reactions is that the same number of electrons should be produced or consumed at each electrode. At equilibrium, ΔrG = 0, the electron flow stops, and oxidation and reduction reactions occur at each electrode with equal rates, as shown schematically in Fig. 1(b) for the case of copper dissolution and deposition. Electrochemical cells in which the electrode reactions take place spontaneously and give rise to an electron flow are called galvanic cells. The chemical energy produced by the cell reaction during the transformation of reactants with high overall free energy into products with low overall free energy can be converted through the electron flow into electrical energy. This principle is used in batteries and fuel cells. The external circuit can be replaced with a direct current (dc) power supply, which will force electrons to go in a direction opposite to the one they tend to go naturally and make the cell reaction proceed in the backward direction to create chemical substances with high overall free energy. This process thus converts electrical into chemical energy and is used in electrolytic cells or battery charging. Corrosion Reactions. The two types of cell processes described previously are of interest when dealing with corrosion. Corrosion reactions are similar to galvanic cells, with the two electrodes being different parts of the metal. Anodic dissolution takes place on certain zones of the surface of a metal and is coupled to cathodic reactions taking place on other zones. The short circuit provided by the conducting metal leads to a continuous metal dissolution. However, application of external potentials, as in electrolytic cells, can be used to protect metals (see the article “Cathodic Protection” in this Volume). While the anodic dissolution involves only the metallic phase, the reduction reaction (cathodic reaction) involves the environment. Several different cathodic reactions (consuming electrons) are encountered in metallic corrosion in aqueous systems. The most common are: •

Proton reduction (in acid media): 2H+(aq) + 2e- → H2

•

Water reduction (in neutral or basic media): 2H2O(l) + 2e- → H2 + 2OH-(aq)

•

Reduction of dissolved oxygen:

•

Metal ion reduction: M3+(aq) + e- → M2+(aq)

•

Metal deposition:

M2+(aq) + 2e- → M Proton reduction in acidic media is very common, and oxygen reduction is also often encountered, because aqueous solutions in contact with air contain significant amounts of dissolved oxygen. Metallic ion reduction and metal deposition are less common and are encountered most often in chemical process streams (Ref 4).

References cited in this section 3. M.G. Fontana, Corrosion Engineering, 3rd ed., McGraw-Hill, 1986, p 448 4. M.G. Fontana, Corrosion Engineering, 2nd ed., McGraw-Hill, 1978, p 251–303


Thermodynamics of Electrochemical Equilibria The Electromotive Force of Galvanic Cells. Consider the following two simple half-cell reactions: oxidation at electrode 1: c1R1 → b1O1 + ne-

(Eq 11)

and reduction at electrode 2: b2O2 + ne- → c2R2

(Eq 12)

where Ri represents a reduced species; Oi is an oxidized species; and bi, ci are stoichiometric coefficients, and n is an integer. It is assumed that the spontaneous galvanic cell reaction is the following: b2O2 + c1R1 → b1O1 + c2R2

(Eq 13)

If the resistance between the electrodes of the galvanic cell is made very high, so that very little current flows, the extent of reaction being small enough not to change the activities of reactants and products, the potential difference between the two electrodes remains constant and is the maximum cell voltage, called the electromotive force (emf) of the cell. The electrical energy produced is then maximum and equal for the transformation of 1 mole of reactants to the absolute value of the molar free energy of the cell chemical reaction: |ΔrG| ≈ Charge passed × Potential difference Conversely, if the cell reaction is driven by an outside source of electric power to go against its spontaneous tendency (as in an electrolytic or charging cell), |ΔrG| is the minimum energy required to drive the backward reaction. Under ideal thermodynamically reversible conditions, where infinitesimal currents are allowed to flow in one or another direction, the cell reaction is at equilibrium, and the perfect equality holds. A noteworthy property is that, whatever the direction of the cell reaction, the polarity of each electrode remains unchanged (see Fig. 1a, Ref 5). If the cell reaction as written in Eq 13 proceeds spontaneously, the electrons flow into an external circuit from electrode 1 where electrons are produced (Eq 11) to electrode 2 where electrons are consumed (Eq 12). Electrode 2 is a cathode, and its polarity is positive. Electrode 1 is an anode, and its polarity is negative. Conversely, if the cell reaction is driven backward by an outside electric power source, the electrons are forced to flow in the opposite direction (from electrode 2 to 1), and half-reactions 1 and 2 are reversed, so electrode 1 becomes a cathode, and electrode 2 becomes an anode, but the polarities of the two

electrodes are left unchanged. Hence, the magnitude and the sign of the cell voltage at equilibrium (emf) depend only on the couple of half-reactions involved. From the thermodynamic convention, the free energy change of a spontaneous cell reaction, which liberates energy, is negative. If the emf is the potential of electrode 2 minus the potential of electrode 1 (ΔE = E2 - E1), and if ΔrG designates the free energy of the cell reaction written in the sense of Eq 13, that is, reduction at electrode 2 (Eq 12) and oxidation at electrode 1 (Eq 11), the relation between ΔrG and ΔE is: ΔrG = -nFΔE

(Eq 14)

where n is the number of electrons exchanged in both half-reactions, and F is the Faraday constant (96,487 coulombs) equal to the charge of 1 mole of electrons. Using Eq 3 and 4, Eq 14 may be rewritten as:

(Eq 15)

Compared to a chemical equilibrium where, at a given temperature, the ratio is equal to a constant Keq (Eq 6), the electrochemical cell at equilibrium is a system with one more degree of freedom, because either the ratio of activities or the cell emf can impose on it. If the cell reaction occurs under conditions in which the reactants and products are in their standard states, the equation becomes: (Eq 16) where ΔE0 is the standard cell emf. The Hydrogen Potential Scale. The absolute potential of an electrode, or even the potential difference between a metal electrode and the surrounding solution, cannot be determined experimentally. It is only possible to measure the voltage across an electrochemical cell, that is, the difference of potential between two identical wires connected to two electrodes. A potential scale may be defined by measuring all electrode potentials with respect to an electrode of constant potential, called the reference electrode. The reference electrode arbitrarily chosen to establish a universal potential scale is the standard hydrogen electrode (SHE). It consists of a platinized platinum electrode (wire or sheet) immersed in an aqueous solution of unit activity of protons, saturated with hydrogen gas at a fugacity of 1 bar. The half-cell reaction is the equilibrium: H+(aq) + e-

H2(g)

(Eq 17)

The SHE possesses the advantages of achieving its equilibrium potential quickly and reproducibly and maintaining it very stable with time (see comparison with other reference electrodes in the article “Potential Measurements with Reference Electrodes” in this Volume). From the convention, the SHE potential is taken as zero. The potential of any electrode can then be determined with respect to this zero reference and is called the potential of the electrode on the standard hydrogen scale, denoted E(SHE). The Potential Sign Convention (Reduction Convention). Before establishing tables of standard potentials for various electrodes on the standard hydrogen scale, it is necessary to fix the convention for the sign of a standard potential value (E0). The potential of any electrode is expressed with respect to the SHE by building (really or virtually) a cell in which the other electrode is a SHE. Consider a typical half-cell reaction with an oxidationreduction (O/R) couple, where O represents the oxidized species and R the reduced species. The cell is represented as: Pt, H2(g)(f = 1 bar)|H+(aq)(a = 1) || O/R

(Eq 18)

Depending on the position of the O/R electrode on the hydrogen scale, the spontaneous single electrode or halfcell reaction will proceed in one direction or the other: R oxidation: cR → bO + ne- combined with proton reduction: H+(aq) + e- → H2(g); or, O reduction: bO + ne- → cR combined with hydrogen oxidation: H2(g) → H+(aq) + e-. For example, the coupling of the oxidation/ reduction couple Fe2+/Fe with the H+/H2 couple brings about the spontaneous oxidation of iron (the free energy of the cell reaction is negative, the Fe2+/Fe electrode is 2+ negative). The situation is entirely different with a Cu /Cu system. If coupled with the H+/H2 couple, the

reduction of copper ions is spontaneous ( , negative; , positive, the polarity of the Cu2+/Cu electrode is positive). The difference between these two metal couples in the direction of the spontaneous reaction with respect to the hydrogen couple must be represented by a sign. This sign will allow cataloging cell potentials from single electrode values and hence compute cell emfs. It was shown previously that the polarity of each electrode in an electrochemical cell is invariant, irrespective of whether the electrode reactions proceed in the spontaneous or the reverse direction. The most meaningful convention is to assign to the potential of the O/R electrode the same sign as its experimentally observed polarity in the cell connecting this electrode to the SHE (Eq 18). The potential of the O/R couple on the standard hydrogen scale, in volts, denoted EO/R(SHE), is thus equal to the cell voltage measured as the potential of the O/R electrode minus the potential of the SHE (zero by convention). When the O/R electrode is in equilibrium, the whole cell is also in equilibrium, and the equilibrium electrode potential versus SHE may be obtained by applying Eq 14 to this particular cell: ΔGred = -nFEO/R(SHE)

(Eq 19)

where ΔGred is the molar free energy change of the balanced reaction of reduction of the oxidized species O by H2(g): (Eq 20) Then:

(Eq 21)

This equilibrium potential, called the reversible potential, is the electrode potential measured under zero current (rest potential) when the electrochemical equilibrium between O and R occurs at the electrode-electrolyte interface. At the International Union of Pure and Applied Chemistry (IUPAC) meeting held in Stockholm in 1953, it was decided that the half-cell reactions should be conventionally written in the reduction direction: bO + ne- ↔ cR

(Eq 22)

and the corresponding equilibrium potentials referred to the SHE obtained from the relation in Eq 21. The reaction as written in Eq 22 stands implicitly for the overall chemical reaction in Eq 20; the electrode equilibrium potential is conveniently written as: (Eq 23) where

is the standard chemical potential or Gibbs free energy of a conventional electron, equal to . It may be checked that: (Eq 24)

The IUPAC convention selects the reduction as the conventional direction for writing electrochemical reactions. This is, of course, not necessarily the spontaneous direction of a reaction, so the corresponding Gibbs free energy change may be positive and the potential negative. As a result of this reduction convention, the potential of the Fe2+/Fe electrode has a negative sign and the one of Cu2+/Cu a positive sign. A negative sign indicates a trend toward corrosion in the presence of H+ ions; that is, the metallic cations have a greater tendency to exist in aqueous solution than the protons. A positive sign indicates, on the contrary, that the proton is more stable than the metallic cations. It is to be noted that there was another convention, called the American convention, where the equation ΔG = -nFE is applied to the reactions written in the oxidation direction. This leads to opposite signs for all oxidation/reduction couples. The advantage of the IUPAC convention, now used worldwide, is that the electrode potential signs correspond to the real polarities measured when the electrode of

interest is connected to the standard H+/H2 electrode and hence are fixed for a given O/R couple, whether reduction or oxidation is considered. The Nernst Equation. If the chemical potentials in Eq 23 are developed using Eq 1, the following expression of the equilibrium or reversible potential for half-cell reaction, known as the Nernst equation, is obtained: (Eq 25)

where aO and aR are the activities of the oxidized and reduced species, respectively, and electrode potential referred to the SHE, corresponding to unit activity of both species:

is the standard

(Eq 26)

Equation 25 shows that the equilibrium potential increases with the ratio of the activities of the oxidized form to the reduced form.

Reference cited in this section 5. L. Pauling, General Chemistry, W.H. Freeman, 1964, p 338–360


Cell Potentials and the Electromotive Force Series The standard equilibrium potentials measured at 25 °C (77 °F) relative to the SHE for various metal-ion electrodes can be tabulated in a series, called the emf series, shown in Table 1. Table 1 Series of standard electrode potentials (electromotive forces) See also (Fig. 2), which shows a schematic of an electrochemical cell used to determine the potential difference of the zinc electrodes versus SHE. Electrode reaction Standard potential at 25 °C (77 °F), volts versus SHE Au3+ + 3e- → Au 1.50 2+ Pd + 2e → Pd 0.987 2+ Hg + 2e → Hg 0.854 Ag+ + e- → Ag 0.800 0.789 + 2e- → 2Hg Cu+ + e- → Cu 0.521 2+ Cu + 2e → Cu 0.337 2H+ + 2e- → H2 0.000 (Reference) Pb2+ + 2e- → Pb -0.126

Sn2+ + 2e- → Sn -0.136 2+ Ni + 2e → Ni -0.250 Co2+ + 2e- → Co -0.277 + T1 + e → T1 -0.336 In3+ + 3e- → In -0.342 2+ Cd + 2e → Cd -0.403 2+ Fe + 2e → Fe -0.440 Ga3+ + 3e- → Ga -0.53 Cr3+ + 3e- → Cr -0.74 2+ Cr + 2e → Cr -0.91 Zn2+ + 2e- → Zn -0.763 Mn2+ + 2e- → Mn -1.18 4+ Zr + 4e → Zr -1.53 Ti2+ + 2e- → Ti -1.63 Al3+ + 3e- → Al -1.66 Hf4+ + 4e- → Hf -1.70 3+ U + 3e → U -1.80 2+ Be + 2e → Be -1.85 Mg2+ + 2e- → Mg -2.37 + Na + e → Na -2.71 2+ Ca + 2e → Ca -2.87 K+ + e- → K -2.93 + Li + e → Li -3.05 For example, the standard electrode potential for zinc—the accepted value for which is -0.763 V/SHE (Table 1)—is obtained by measuring the emf of a cell made of a SHE and a zinc electrode in a zinc salt solution of unit activity (Fig. 2). As described previously, the emf is a measurement of the maximum potential that exists in the galvanic cell when zero current is flowing between the electrode and the SHE.

Fig. 2 Electrochemical cell containing a standard zinc electrode and a standard hydrogen electrode (SHE) (H2 fugacity = 1 bar). The measurement of the cell voltage gives the standard equilibrium potential of the Zn2+/Zn couple versus SHE. This procedure can be repeated by exchanging the zinc electrode with any other metal, and the surrounding electrolyte by a solution of metal ion, in order to obtain the series of standard half- cell potentials listed in Table 1. The position of a particular metal in the emf series gives an indication of the tendency of the metal to be oxidized. The higher the metal is on the potential scale, the more noble it is and resistant to oxidation; the lower

on the scale it is, the more easily oxidized it is. The metals listed in Table 1 below the H+/H2 couple (SHE) are easily oxidized and reduce protons into H2(g), while the metals listed above the SHE have their metal ions reduced into metal by H2(g). However, care should be taken when predicting the behavior of a metal from the data in Table 1, because they are given for standard conditions, usually not close to the real corrosion conditions. Changes in concentration (i.e., activity) or temperature will change the electrode potentials according to Eq 25. The comparison between the standard potentials of half-cells from Table 1 is used to predict the abilities of metals to reduce other metal ions from solution. For example, consider the Daniell cell, constituted as follows: (Eq 27) 2+

The potential of the Cu /Cu couple is higher than the potential of the Zn2+/Zn couple. If the electrodes are coupled together in a galvanic cell, electrons will flow from the zinc electrode, which produces electrons via the oxidation reaction, through the external circuit to the copper electrode, which consumes electrons via the reduction reaction. The calculated standard emf for this cell is ΔE0 = E0(Cu+2/Cu) - E0(Zn+2/Zn) = +0.337 - (0.763) = +1.100 V/SHE. Then, the standard free energy for the cell reaction is (Eq 16) ΔrG0 = -2FΔE0 = -212.3 kJ/mol. Its sign indicates that there is a natural tendency for zinc atoms to reduce cupric ions and be oxidized into Zn2+ (Eq 8). It is important to note that the calculated equilibrium potentials are for a very specific condition (standard state) and may not apply to a specific corrosion environment; also, there are thermodynamic quantities that do not take into account the factors that may limit a reaction, such as slow kinetics or protection by corrosion product layers. More complete emf series (Ref 6, 7) and potentials in other environments (Ref 8) are available.

References cited in this section 6. A.J. Bard, R. Parsons, and J. Jordan, Standard Potentials in Aqueous Solutions, Marcel Dekker, 1985 7. W.M. Latimer, Oxidation Potentials, Prentice-Hall, 1964 8. F.L. La Que, Corrosion Handbook, H.H. Uhlig, Ed., John Wiley & Sons, 1948, p 416


Acknowledgment Portions of this article have been adapted from C.A. Natalie, Electrode Potentials, Corrosion, Vol 13, Metals Handbook, 9th ed., ASM International, 1987, p 19–21.


References 1. J.M. Smith and H.C. Van Hess, Introduction to Chemical Engineering Thermodynamics, McGraw-Hill, 1975, p 159–162 2. K. Denbigh, Principles of Chemical Equilibrium, 2nd ed., Cambridge Press, 1981, p 133–186 3. M.G. Fontana, Corrosion Engineering, 3rd ed., McGraw-Hill, 1986, p 448 4. M.G. Fontana, Corrosion Engineering, 2nd ed., McGraw-Hill, 1978, p 251–303 5. L. Pauling, General Chemistry, W.H. Freeman, 1964, p 338–360 6. A.J. Bard, R. Parsons, and J. Jordan, Standard Potentials in Aqueous Solutions, Marcel Dekker, 1985 7. W.M. Latimer, Oxidation Potentials, Prentice-Hall, 1964 8. F.L. La Que, Corrosion Handbook, H.H. Uhlig, Ed., John Wiley & Sons, 1948, p 416


Selected References •

• • •

J.W. Cobble, R.C. Murray, Jr., P.J. Turner and K. Chen, High-Temperature Thermodynamic Data for Species in Aqueous Solution, San Diego State University Foundation, Electric Power Research Institute (EPRI) Report NP- 2400, EPRI, May 1982 K.C. Mills, Thermodynamic Data for Inorganic Sulfides, Selenides and Tellurides, Butterworths, 1974 D.R. Stull and H. Prophet, JANAF Thermochemical Tables, National Bureau of Standards, NSRDSNBS 37, U.S. Government Printing Office, Washington, D.C., 1971 D.D. Wagman, W.H. Evans, V.B. Parker, R.H. Schumm, I. Halow, S.M. Bailey, K.L. Churney, and R.L. Nuttall, J. Phys. Chem. Ref. Data, Vol 11 (Supplement 2), 1982

E. Protopopoff and P. Marcus, Potential Measurements with Reference Electrodes, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 13–-16

Potential Measurements with Reference Electrodes E. Protopopoff, Laboratoire de Physico-Chimie des Surfaces, CNRS, and P. Marcus, Ecole Nationale Supérieure de Chimie de Paris, Université Pierre et Marie Curie

Introduction ELECTRODE POTENTIAL MEASUREMENT is an important aspect of corrosion studies and corrosion prevention. It is included in any determination of the corrosion rate of metals and alloys in various environments and in the control of the potential in cathodic and anodic protection. The potential of an electrode can be determined only by measuring the voltage in an electrochemical cell between this electrode and an electrode of constant potential, called the reference electrode. Many errors and problems can be avoided by careful selection of the best reference electrode for a specific case and by knowledge of the electrochemical principles that control the potential measurements in order to obtain meaningful measurements. A reference electrode, once selected, must be properly used, taking into account the stability of its potential and the problem of ohmic (IR) drop. Many different reference electrodes are available, and others can be designed by the users themselves for particular situations. Each electrode has its characteristic rest potential value, which is used to convert potential values measured with respect to this reference into values expressed with respect to other references. In particular, the conversion of the potentials from or to the hydrogen scale is frequently required for use of potential- pH diagrams, which are discussed in the article “Potential versus pH (Pourbaix) Diagrams” in this Section of the volume.

E. Protopopoff and P. Marcus, Potential Measurements with Reference Electrodes, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 13–-16 Potential Measurements with Reference Electrodes E. Protopopoff, Laboratoire de Physico-Chimie des Surfaces, CNRS, and P. Marcus, Ecole Nationale Supérieure de Chimie de Paris, Université Pierre et Marie Curie

The Three-Electrode Device When a system is at rest and no significant current is flowing, the use of only one other electrode as a reference is sufficient to measure the test (or working) electrode potential versus the reference potential. When a current is flowing spontaneously in a galvanic cell or is imposed on an electrolytic cell, reactions at both electrodes are not at equilibrium, and there is consequently an overpotential on each of them. The potential difference measured between the two electrodes then includes the value of the two overpotentials, and it is not possible to determine the potential of the test electrode. To obtain this value, a third electrode, the auxiliary or counterelectrode, must be used (Fig. 1). In this arrangement, current flows only between the test and the auxiliary electrodes. A high-impedance voltmeter placed between the test and the reference electrodes prevents any significant current flow through the reference electrode, which then shows a negligible overpotential and remains very close to its rest potential. Most reference electrodes can be damaged by current flow. The test electrode potential and its changes under electric current flow can then be measured with respect to a fixed reference potential. The three- electrode system is widely used in laboratory and field potential measurement.

Fig. 1 Three-electrode device. V, voltmeter


Electrode Selection Characteristics A good reference electrode must reach its potential quickly, be reproducible, and remain stable with time. It must have a practically nonpolarizable metal-solution interface; that is, its potential must not depart significantly from the equilibrium value on the passage of a small current across the interface. The potential of the junction between the electrolytes of the reference and test electrodes must be minimized. These criteria are detailed subsequently. Stable and Reproducible Potential. Electrodes used as references should rapidly achieve a stable and reproducible potential that is free of significant fluctuations. To obtain these characteristics, it is advantageous, whenever possible, to use reversible electrodes, which can easily be made. The reference electrode arbitrarily chosen to establish a universal potential scale is the standard hydrogen electrode (SHE). It consists of a platinized or black platinum wire or sheet immersed in an aqueous solution of unit activity of protons saturated with hydrogen gas at a fugacity of one bar (14.5 psia). The half-cell reaction is H+(aq) + e- ↔ H2(g). Any non-standard reversible hydrogen electrode with well-controlled H+ activity and H2 fugacity can also be used as a reference. The equilibrium potential (Eeq) of a nonstandard reversible hydrogen electrode versus the SHE is, from the Nernst equation (Eq 25 of the article “Electrode Potentials” in this Section of the volume): (Eq 1)

where R is the gas constant, T is the absolute temperature, F is the Faraday constant,

is the proton activity

in solution, and is the H2 fugacity near the electrode; the SHE potential, , is, by convention, equal to zero. Platinization of a smooth platinum electrode is achieved by electrodeposition from a solution of H2PtCl6 of a black platinum layer having a very rough surface and hence a very high specific area. The electrolyte is made up of a mineral acid (HCl or H2SO4) with a well-defined activity of H+. Hydrogen gas is bubbled on the black platinum. While the hydrogen electrode is the fundamental reference, it has certain disadvantages in real conditions: the electrolyte must be prepared with an accurately known proton activity; the hydrogen gas must be purified, particularly from oxygen; furthermore, the platinum electrode must be frequently replatinized, because it easily gets “poisoned” by the adsorption of impurities present in the solution, which prevents the establishment of the equilibrium potential. For these reasons, practical corrosion measurements are usually not performed with the SHE but with secondary reference electrodes that are easier to construct and handle, less sensitive to impurities, and whose potential is very stable and well- known with respect to the SHE (Ref 1). The simpler reference electrodes are metal-ion (Mz+/M) electrodes, also called metallic electrodes of the first kind. The copper-copper sulfate (CuSO4/Cu) electrode is an excellent example of a good reversible electrode and it is widely used as a reference electrode in the corrosion field. It can easily be made by immersing a copper wire in a glass tube filled with a CuSO4 aqueous solution and terminated by a porous plug (to allow ionic conduction with the cell electrolyte), as shown in Fig. 2.

Fig. 2 Schematic of a copper/copper sulfate reference electrode This electrode is reversible, because a small cathodic current produces the reduction reaction (Cu2+ + 2e- → Cu), while an anodic current brings about the oxidation reaction (Cu → Cu2+ + 2e-). Copper is a semi-noble metal and does not dissolve anodically in a solution of protons. In the case of the CuSO4/Cu electrode, the rest potential is equal to the equilibrium potential that can be computed from the Nernst equation: (Eq 2) where the standard potential Potentials” in this Volume), and

= +0.337 V versus SHE (see Table 1 in the article “Electrode is the activity of Cu2+ in the aqueous solution. If a copper solution of

concentration 1.00 mol/L is used (where < 1), the equilibrium potential of the CuSO4/Cu electrode takes the value +0.310 V versus SHE at 25 °C (77 °F). This well-defined reversible electrode is reliable and easy to build. Another common metal-ion electrode is the silver electrode (Ag+/Ag). The Nernst equation applied to the halfcell reaction Ag+ + e- ↔ Ag gives: (Eq 3) where the standard potential = 0.800 V versus SHE. Metallic electrodes of the second kind (also called secondary reference electrodes) are similar electrodes where the potential-determining activity of the metallic ion, , in solution is controlled by putting the metal M in contact with a sparingly soluble M compound (salt, oxide, hydroxide), itself in contact with the solution. Then, is determined by the solubility product of the salt and the activity of the anion. For example, if silver chloride (AgCl), only slightly soluble in water, is present on the silver surface, then the following equilibrium holds: AgCl ↔ Ag+ + Cl-

(Eq 4)

= Ks(AgCl). Expressing from this The equilibrium mass law gives the solubility product as relation and placing it in Eq 3 leads to the following expression of the equilibrium potential of the Ag+, Cl/AgCl/Ag half-cell:

(Eq 5)

corresponding to the global electrode equilibrium: AgCl + e- ↔ Ag + Cl-

(Eq 6)

The standard potential of this reversible silver-silver chloride (Ag/Cl) electrode is + ln Ks(AgCl). At 25 °C, (77 °F), because

=

= 0.800 V/SHE and Ks(AgCl) = 1.78 · 10-10,

= +0.222 V/SHE (Ref 2). From Eq 5, the potential of this type of reference electrode depends on the activity of the anion in solution (Cl- here), and this activity is controlled by the addition of a soluble salt of this anion (KCl). If a KCl solution of concentration 1.00 mol/L is used (where < 1), the equilibrium potential of the AgCl/Ag electrode takes the value +0.237 V/ SHE at 25 °C (77 °F). In a saturated KCl solution (where > 1), it is +0.198 V/SHE (Ref 2). Figure 3 shows a typical Ag-AgCl electrode. The AgCl coating is made by anodization of the silver wire in a chloride-containing solution. This electrode is easily assembled and can be placed directly into the electrochemical cell. Quite similar are the Pb/PbSO4 and the Ag-Ag2O (Ref 1, 2).

Fig. 3 Schematic of silver-silver chloride and calomel reference electrodes. Source: Ref 12 The most used secondary reference electrode is the calomel electrode (Hg-Hg2Cl2). The sparingly soluble salt is, in this case, calomel (Hg2Cl2) which floats as a paste on the top of a liquid mercury drop and which, in contact with a KCl solution, dissociates slightly into AgCl electrode, the following relation is obtained:

and Cl- ions. Using the same method as for the Ag-

(Eq 7)

corresponding to the global electrode equilibrium: Hg2Cl2 + 2e- ↔ 2Hg + 2Cl-

(Eq 8)

Figure 3 shows a typical calomel electrode. A platinum wire connects the electrode to the rest of the circuit. The most usual calomel electrodes are prepared with KCl solutions at a unit molarity of Cl- (normal calomel electrode or NCE) or saturated calomel electrode or (SCE). At 25 °C (77 °F), = 0.268 V/SHE, ENCE = 0.281 V, and ESCE = 0.242 V/SHE (Ref 2). The SCE has the advantage of being the easiest to prepare. However, due to the high temperature dependence of the KCl solubility, its potential varies markedly with temperature (~1 mV/°C as compared to 0.1 mV/°C for the NCE) (Ref 2), so its use makes mandatory the accurate monitoring of the temperature of experiment. Moreover, one must be careful when using a calomel electrode that only very low currents pass through the interface, because HgO may form, which irreversibly spoils the electrode.

Similar reference electrodes for measurements in aqueous solutions are the mercurous sulfate electrode ( , ) in a solution of potassium sulfate, and the mercuric oxide electrode (Hg2+, OH/HgO/Hg) in a solution of sodium hydroxide. Their potentials at 25 °C (77 °F) may be found in the literature (Ref 1, 2). All of these secondary reference electrodes are reversible. In some practical cases, nonreversible electrodes such as graphite are used. Although not as good, their potential stability in a particular environment is considered sufficient for certain applications. In the selection of reference electrodes, their durability and price must also be considered. Low Polarizability. A good reference electrode must have a practically nonpolarizable metal-solution interface; that is, its potential must not depart significantly from the equilibrium value on the passage across it of a small current (even minimized in the three-electrode device, Fig. 1), because the electrode polarization introduces an error in the potential measurement. The potential versus current density response of a good reference electrode, called a polarization curve, should be as flat as possible. Consider that the electrode equilibrium is bO + ne- = cR, where O is the oxidized species; R is the reduced species; and n is an integer. The net current density across the interface is proportional to the difference between the anodic and the cathodic rates. If the electrode reaction rate is limited by the electron transfer across the metal-solution interface (electron transfer is slow compared to mass transport of O and R between the bulk solution and the interface), the net current density is simply related to the overpotential η = E - Eeq, where E is the potential, and Eeq is the equilibrium potential, by (Ref 3): (Eq 9) where i0 is the exchange current density of the electrode reaction; and α is the charge transfer coefficient for the anodic reaction (0 < α < 1), whose value is close to 0.5 for a single-step reaction (n = 1). (More detailed information on polarization curves can be found in the articles “Kinetics of Aqueous Corrosion” and “Electrochemical Methods of Corrosion Testing” in this Volume). For potentials close to the equilibrium potential, such as η < RT/F (26 mV at room temperature), the relation (Eq 9) can be approximated by a linear i versus η dependence: (Eq 10) or (Eq 11) The term RT/nFi0 = (dη/di)0 is called the polarization resistance, Rp. A good reference electrode should have a low polarization resistance, which implies high exchange current density. This happens for high rate constants for the anodic and cathodic reactions and high concentrations of reacting species O and R. With regard to this criterion, the SHE that has i0 > 10-3 A/ geometric cm2 is a particularly good reference electrode (Ref 3). It is a question of judgment how polarizable the reference electrode can be. The answer depends on the precision required and the impedance of the voltmeter used. A high-impedance voltmeter (1012 ohms) may provide acceptable results with a relatively polarizable electrode. The Liquid Junction Potential. Reference electrodes are usually made of a metal immersed in a well-defined electrolyte. In the case of the CuSO4/Cu electrode, the electrolyte is a saturated CuSO 4 aqueous solution; for the SCE, it is a saturated KCl solution. This electrolyte that characterizes the reference electrode must come into contact with the liquid environment of the test electrode to complete the measuring circuit (Fig. 4). There is direct contact between different aqueous media. The difference in chemical composition of the two solutions produces a phenomenon of interdiffusion. In this process, except for a few electrolytes such as KCl, the cations and anions move at different speeds. As an example, in hydrogen chloride (HCl) solution in contact with another medium, the H+ ions move faster than the Cl- ions. As a result, a charge separation appears at the limit between the two liquids, the liquid junction. This produces a potential difference called the liquid junction potential, which is included in the measured voltage, V, as expressed in:

V = VT - VR + VLJP

(Eq 12)

where VT is the test potential to be measured, VR is the reference electrode potential, and VLJP is the unknown liquid junction potential.

Fig. 4 Schematic of an electrochemical cell with liquid junction potential. P, interface; V, voltmeter In order to determine VT, the liquid junction potential must be eliminated or minimized. The best way, when possible, is to design a reference electrode using an electrolyte identical to the solution in which the test electrode is immersed (Fig. 4). However, in most cases this is not possible, and the best approach is to minimize the liquid junction potential by using a reference electrolyte with a chemical composition as close as possible to the corrosion environment. The use of a solution of KCl (such as in the calomel electrode) offers a partial answer. The diffusion rates of potassium (K+) and chloride (Cl-) ions are similar. In contact with another electrolyte, a KCl solution does not produce much charge separation and, consequently, no significant liquid junction potential. The ions present in the other solution, however, also diffuse, and they may do so at different rates, thus producing some separation of charge at the interface (P in Fig. 4). The remaining liquid junction potential, after minimization, constitutes an error that is frequently accepted in electrode potential measurements, especially when compared with results determined under similar experimental conditions. While liquid junction potentials must be minimized as much as possible, there is no general solution for this; each individual case must be well thought out.

References cited in this section 1. D.J.G. Ives and J. Janz, Reference Electrodes, Academic Press, 1961 2. C.H. Hamann, A.H. Hamnett, W. Vielstich, Electrochemistry, Wiley-VCH, Weinheim, 1998 3. M.G. Fontana, Corrosion Engineering, 2nd ed., McGraw-Hill, 1978, p 12


Operating Conditions for Reference Electrodes When a reference electrode has been selected for a particular application, its proper use requires caution and specific measurement conditions. When measuring the potential of a polarized test electrode versus a reference electrode, it is important not to polarize or damage the latter by applying a significant current density. Also, the ohmic (IR) drop must be minimized. Very Low Current Density. It is important to use a reference electrode that operates at its known open-circuit potential and thus avoid applying any significant overpotential to it. This is achieved by using a highimpedance voltmeter that has a negligible input current and, for test electrode polarization measurements, by using an auxiliary electrode in a three-electrode system (Fig. 1). The value tolerated for the maximum overpotential on the reference electrode, at the condition that it stays under the limit over which the electrode suffers irreversible damages (like the calomel electrode), is a matter of judgment that depends on the accepted magnitude of error in the particular case under investigation. The use of an electrometer or a high-impedance voltmeter (1012 ohms) fulfills the usual requirements. When a lower impedance instrument is used, an unacceptable overpotential could result if the electrode is too polarizable. The IR Drop and Its Mitigation. The IR drop is an ohmic voltage that results from electric current flow in ionic solutions. Electrolytes have an ohmic resistance; when a current passes through them, an IR voltage can be observed between two distinct points. When the reference electrode is immersed at some distance from a working or test electrode, it is in the electric field somewhere along the current path. An electrolyte resistance exists along the path between the test and the reference electrodes. As current flows through that path, an IR voltage appears in the potential measurement according to: V = VT - VR + IR

(Eq 13)

where VT is the test potential to be measured, VR is the reference electrode potential, and IR is the ohmic drop. In this case, the liquid junction potential has been neglected. The IR drop constitutes a second unknown value in a single equation. It must be eliminated or minimized. The Luggin capillary is a tube, usually made of glass, that has been narrowed by elongation at one end. The narrow end is placed as close as possible to the test electrode surface (Fig. 1), and the other end of the tube goes to the reference electrode compartment. The Luggin capillary is filled with cell electrolyte, which provides an electric link between the reference and the test electrode. The use of a high-impedance voltmeter prevents significant current flow into the reference electrode and into the capillary tube between the test electrode and the reference electrode compartment (Fig. 1). This absence of current eliminates the IR drop, and the measurement of VT is then possible. A residual IR drop may, however, exist between the tip of the Luggin capillary and the test electrode. This is usually negligible, however, especially in high-conductivity media. The remote electrode technique can be used only for measurement in an electrolyte with very low resistivity, usually in the laboratory. It is applicable, for example, in a molten salt solution, in which the ohmic resistance R is very small. In such a case, the reference electrode can be placed a few centimeters away from the test electrode, because the IR drop remains negligible. In other electrolytes (for example, in measurements in soils), the ohmic resistance is rather large, and the IR drop cannot be eliminated in this manner. The Current Interruption Technique. In this case, when the current is flowing, the IR drop is included in the measurement. A recording of the potential is shown in Fig. 5. At time t1, the current is interrupted so that I = 0 and IR = 0.

Fig. 5 The potential decay at current interruption. IR is the potential drop due to the electrolyte ohmic resistance. At the moment of the interruption, however, the electrode is still polarized, as can be seen at point P in Fig. 5. The progressive capacitance discharge and depolarization of the test electrode take some time. The potential measured at the instant of interruption then represents the test electrode potential corrected for the IR drop. Precise measurements of this potential are obtained with an oscilloscope. Potential Conversion Between Reference Electrodes. Due to the number of different reference electrodes used, each potential measurement must be accompanied by a clear statement of the reference used. It is often needed to express electrode potentials versus a particular reference, regardless of the actual reference used in the measurement. The procedure is illustrated in the following example. The electrode potential of a buried steel pipe is measured with respect to a CuSO4/Cu electrode, and the value is -650 mV for a pH 4 environment. If that value is mistakenly placed in the iron E-pH (Pourbaix) diagram (Fig. 1 in the article “Potential versus pH (Pourbaix) Diagrams” in this Volume), it could be concluded that corrosion will not occur. This conclusion, however, would be incorrect, because the E-pH (Pourbaix) diagrams are always computed with respect to the SHE. It is then necessary to express the measured electrode potential with respect to the SHE before consulting the E-pH (Pourbaix) diagram. The CuSO4/Cu electrode potential is +310 mV versus SHE, so this value must be added to the measured potential: ESHE = -650 + 310 = -340 mV. The principle of this conversion is illustrated in the electrode potential conversion diagram of Fig. 6.

Fig. 6 Diagram of potential conversion between reference electrodes. SHE, standard hydrogen electrode; CuSO4, copper-copper sulfate electrode. SHE, standard hydrogen electrode

The value of -340 mV placed in the E-pH Pourbaix diagram at a pH 4 clearly lies in the corrosion region for iron. It would then be definitely necessary to consider the cost benefit of a protection system for the steel pipe.


Acknowledgment Portions of this article have been adapted from D.L Piron, Potential Measurements with Reference Electrodes, Corrosion, Vol 13, Metals Handbook, 9th ed., ASM International, 1987, p 21–24.


References 1. D.J.G. Ives and J. Janz, Reference Electrodes, Academic Press, 1961 2. C.H. Hamann, A.H. Hamnett, W. Vielstich, Electrochemistry, Wiley-VCH, Weinheim, 1998 3. M.G. Fontana, Corrosion Engineering, 2nd ed., McGraw-Hill, 1978, p 12 4. L. Pauling, General Chemistry, W.H. Freeman, 1964, p 338–360

E. Protopopoff and P. Marcus, Potential versus pH (Pourbaix) Diagrams, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 17–-30

Potential versus pH (Pourbaix) Diagrams E. Protopopoff and P. Marcus, CNRS, Ecole Nationale Supérieure de Chimie de Paris, Université Pierre et Marie Curie

Introduction THE PRINCIPLE OF POTENTIAL-pH DIAGRAMS was established in the 1940s in Belgium by Marcel Pourbaix (Ref 1, 2, 3, 4). A potential- pH diagram is a graphical representation of the relations, derived from the Nernst equation, between the pH and the equilibrium potentials (E) of the most probable electrochemical reactions occurring in a solution containing a specific element. The standard equilibrium potentials are computed from thermodynamic data (standard chemical potentials, or Gibbs free energies of formation). The equilibrium relations drawn for a given concentration of the element or for a given ratio of activities of two dissolved species of the element give E-pH lines. The representation of the equilibrium pHs for acid-base reactions (independent of the potential) gives vertical lines. All those lines delimit E-pH domains of stability for the various species of the element, metal, ions, oxides, and hydroxides. Potential- pH diagrams synthesize many important types of information that are useful in corrosion and in other fields. They make it possible to discern at a glance the stable species for specific conditions of potential and pH (Ref 1, 2, 3, 4). The principle of E-pH diagrams may be simply understood with the case of iron in water. Corrosion in deaerated water is expressed by the electrochemical reaction Fe → Fe2+ + 2e-. The equilibrium potential for the Fe2+/Fe couple can be calculated using the Nernst equation: (Eq 1) where is the standard potential value for the couple, R is the gas constant, T is the absolute temperature, F is the Faraday constant, and is activity for the ferrous ion in solution. For a given temperature and Fe2+ concentration (activity ), the equilibrium potential is constant and is represented as a horizontal line in a E-pH diagram (Fig. 1). This line indicates the potential at which Fe and Fe2+ at a given concentration are in equilibrium and can coexist with no net tendency for one to transform into the other. At potentials above the line, iron metal is not stable and tends to dissolve as Fe2+, hence the Fe2+ concentration increases until a new equilibrium is reached; this is a domain of stability for Fe2+. At potentials below the equilibrium line, the stability of the metallic iron increases, Fe2+ tends to be reduced, and thus its concentration decreases; this is the domain of stability for the metal (Fig. 1).

Fig. 1 Iron E-pH diagram. Dashed lines a and b are explained in Fig. 7 and in the corresponding text. The diagrams of all metal-water systems have the same common features; the lower E-pH lines give the limit between the domain of stability of the metal and the domain of stability of either the first metallic ion or the first metallic oxide. For E-pH conditions below these lines, the metal is stable, and corrosion cannot take place. This is the immunity region (Fig. 1). For E-pH conditions above the line for the equilibrium between the metal and the first metallic ion, the metal is not stable, and it tends to be oxidized and dissolved into ions. The system is then in the corrosion or activity region of the diagram. Besides the main corrosion region in the stability domains of the metallic ions at low pH (acid corrosion), there is generally also a smaller domain of stability of oxygenated metallic ions at high pHs, leading to alkaline corrosion (Fig. 1). When the reaction of the metal with water produces an oxide (or hydroxide) that forms a protective layer, the metal is said to be passivated. For E-pH conditions above the lines for the metal-oxide and ion-oxide equilibria, the system is in the passivation region (Fig. 1). Diagrams such as Fig. 1, (Ref 1, 2, 3, 4), have proved to be useful in corrosion as well as in many other fields, such as industrial electrolysis, plating, electrowinning and electrorefining of metals, primary and secondary

electric cells, water treatment, and hydrometallurgy. It is important to emphasize that these diagrams are based on thermodynamic calculations for a number of selected chemical species and the possible equilibria between them. It is possible to predict from a E-pH diagram if a metal will tend to corrode or not. It is not possible, however, to determine from these diagrams alone how long a metal will resist corrosion. Pourbaix diagrams offer a framework for kinetic interpretation, but they do not provide information on corrosion rates (Ref 3). They are not a substitute for kinetic studies. Each E-pH diagram is computed for selected chemical species corresponding to the possible forms of the element considered in the solution under study. The addition of one or more elements, for example, carbon, sulfur, or chlorine, to a system will introduce new equilibria. Their representation in the E-pH diagram will produce a new diagram more complex than the previous one.

References cited in this section 1. M. Pourbaix, Thermodynamique des Solutions Aqueuses Diluées, Potentiel D'oxydo- Réduction (résumé de conférence), Bull. Soc. Chim. Belgique, Vol 48, Dec 1938 2. M. Pourbaix, Thermodynamics of Dilute Aqueous Solutions, Arnold Publications, 1949 3. R.W. Staehle, Marcel J.N. Pourbaix—Palladium Award Medalist, J. Eletrochem. Soc., Vol 123, 1976, p 23C 4. M. Pourbaix, Atlas of the Electrochemical Equilibria, NACE, 1974

E. Protopopoff and P. Marcus, Potential versus pH (Pourbaix) Diagrams, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 17–-30 Potential versus pH (Pourbaix) Diagrams E. Protopopoff and P. Marcus, CNRS, Ecole Nationale Supérieure de Chimie de Paris, Université Pierre et Marie Curie

Calculation and Construction of E-pH Diagrams Potential-pH diagrams are based on thermodynamic calculations. The equilibrium lines that set the limits between the various stability domains are calculated for the various electrochemical or chemical equilibria between the chemical species considered. There are three types of reactions to be considered: • • •

Electrochemical reactions of pure charge (electron) transfer Electrochemical reactions involving both electron and solvated proton (H+) transfer Acid-base reactions of pure H+ transfer (no electrons involved)

Pure Charge Transfer Reactions. These electrochemical reactions involve only a reduced species on one side and an oxidized species and electrons on the other side. They have no solvated protons (H+) as reacting particles; consequently, they are not influenced by pH. An example of a reaction of this type is the oxidation/ reduction of Ni/Ni2+: Ni Ni2+ + 2e-. From the Nernst equation (Eq 1), the equilibrium potential for the 2+ couple Ni /Ni can be written:

(Eq 2)

where is the standard potential for the couple and is the activity of Ni2+ in the solution; R, T, and F are defined in Eq 1; at a temperature of 25 °C (298.15 K), RT ln 10/F ≈ 0.059 V. The potential depends on the Ni2+ activity but not on H+ ions. It is then independent of the pH. For a given value of , it is represented by a horizontal line in a E-pH diagram. The standard potential, Eo, is: (Eq 3) By convention, at 25 °C (298.15 K), the standard chemical potentials of a species under elemental form and of the conventional electron are equal to 0. This gives

and simplifies the previous equation: (Eq 4)

The values of the standard chemical potentials at 25 °C are found in the Atlas of Electrochemical Equilibria (Ref 4) or listed as standard Gibbs free energies of formation in tables of chemical thermodynamic properties (Ref 5) or in specialized papers. They are given for the formation of the substance from its elements in their standard states. The standard state for the dissolved species is the hypothetical ideal solution of the substance at unit molality (number of moles of the species per kilogram of water). At 25 °C, the numerical value of the molality may be approximated by the value of the molarity in number of moles per liter of water. at 25 °C is −48.2 kJ/mol (Ref 4), the standard potential is:

If the value of

(Eq 5) This result can be introduced into Eq 2 to give the equilibrium potential for the couple Ni2+/Ni at 25 °C: (Eq 6) As previously stated, this potential depends only on the activity of Ni2+, not on the pH. It is usually assumed in E-pH diagrams that the solution behaves ideally; that is, the activity coefficient of each dissolved species is unity, so the activity of the species is simply equal to the numerical value of its molality (or molarity at 25 °C). This assumption is reasonably accurate for molalities of less than 10-3 mol/kg in the pH range 2 to 12. Outside these limits, the Pourbaix diagram is valid only for activities, and molalities must be corrected. It is customary to select four molalities (molarities at 25 °C): 1 or 100, 10-2, 10-4, and 10-6 mol/kg (mol/L) to show the effect of ion concentration. This will provide four horizontal lines, as shown in Fig. 2: •

At an activity of 10o,

•

With 10-2,

= −0.25 + 0.030 log 10-2 = -0.31 V

•

With 10-4,

= −0.37 V

•

-6

With 10 ,

= −0.25 V

= −0.43 V

Fig. 2 Partial E-pH diagram for the equilibrium Ni2+ + 2e- = Ni for various values of log For any activity of Ni2+ in the solution, a horizontal line represents the potential at which nickel ions at the given activity and nickel metal are in equilibrium. Above the line is the region of stability of Ni2+ ions at a higher activity; nickel metal at these potentials will tend to corrode (dissolve) and produce Ni 2+, the stable species. Below the line, Ni2+ ions will tend to be reduced into metallic nickel, and nickel in this condition is stable and will not corrode (dissolve). It is immune to corrosion. Reactions Involving Both Electron and Proton Transfer. Consider a typical electrochemical reaction with proton transfer in an aqueous electrolyte: bO + mH+ + ne-

cR + dH2O

(Eq 7)

where O and R are the oxidized and the reduced form of the considered element, respectively; b, m, c, and d are stoichiometric coefficients; and n is an integer. From the Nernst equation, the equilibrium potential is equal to: (Eq 8) with Δox = = − ; where is the standard Gibbs free energy charge of reduction. The activity of water can be taken equal to unity in not too concentrated aqueous solutions. If O or R is a pure solid phase (oxide, hydroxide, or metal), aO or aR is equal to unity. The pH of the solution is pH = -log aH+. Hence:

(Eq 9)

This equation shows that the equilibrium potential corresponding to a given ratio r = (aO)b/ (aR)c is a linear function of pH (decreasing if m is positive). As an example, nickel can react with water to form an oxide, according to the electrochemical reaction: Ni + H2O

NiO + 2H+ + 2e-

(Eq 10)

The Nernst equation can be written as follows:

(Eq 11) where the standard potential Eo is given by:

From the data in Ref 4:

The NiO and Ni are solid phases, and they are considered to be pure; their activity is therefore 1. Equation 11 can then be simplified, and the equilibrium potential at 25 °C becomes: Eeq(NiO/Ni)(298.15 K) = +0.11 - 0.059 pH

(Eq 12)

In this case, the equilibrium potential decreases with an increase in pH, as represented in the partial E-pH diagram of Fig. 3. The diagonal line gives the value of the equilibrium potential of the NiO/Ni couple at all pH values. Above the line, NiO is stable, and below it, nickel metal is stable.

Fig. 3 Partial E-pH diagram for the equilibrium Ni + H2O = NiO + 2H+ + 2ePotential-pH diagrams are very general and can also be applied to electrochemical reactions involving nonmetallic elements. An example involving the reduction of nitrite ion ( ) to ammonium ion ( ) is given here. In this case, the metal of the electrode supports the reaction by giving or taking away electrons, as follows: The Nernst equation gives:

(Eq 13)

where the standard potential is expressed as:

From the data in Ref 4:

The equilibrium potential at 25 °C is then given by: (Eq 14) This E-pH relation is represented for equal activity in

Fig. 4 Partial E-pH diagram for the equilibrium Above the line, the ratio

and

by the line drawn in Fig. 4.

+ 8H+ + 6e- =

+ 2H2O

is higher than unity, so this is a region where

is

predominantly stable. Below the line, the ratio is lower than unity, so this is a region where is predominantly stable. Acid-Base Reactions. Consider a typical acid-base chemical reaction in aqueous solution, between the acid form C and the basic form B: cC + dH2O = bB + mH+

(Eq 15)

When the reaction is in thermodynamic equilibrium:

where ΔrG0 is the standard Gibbs free energy change, μ0 represents the standard chemical potentials of the different substances, and Keq is the equilibrium constant for the reaction: Keq = . The activity of water may be taken equal to unity in not too concentrated aqueous solutions. If B or C is a pure solid phase (metal, oxide, or hydroxide), aB or aC is equal to unity. Using pH = , the previous equation may be rearranged as: (Eq 16) This equation shows that the ratio of the activities of the basic form to the acid form increases with pH. The pH at equilibrium corresponding to a given ratio r = (aB)b/(aC)c can be determined from:

(Eq 17) with pKeq = -log Keq = ΔrG0/2.303 RT. This equilibrium pH does not depend on the potential and will be represented on the E-pH diagram by a vertical line. As an example, in the case of cobalt, Co2+ and CoO are involved in an acid-base reaction: Co2+ + H2O

CoO + 2H+

(Eq 18)

CoO and H2O both have activities of 1, so the application of Eq 17 gives: (Eq 19) with:

By replacing the standard chemical potentials by their values given in the Atlas of Electrochemical Equilibria (Ref 4):

and finally:

Thus, for an activity unity for Co2+, pHeq = 6.3. This value can also be obtained by the intersection of two equilibrium E-pH lines. As shown previously for nickel, there are electrochemical equilibria between the metal and its first ion (Co2+) and between the metal and its first oxide (CoO). It is possible to determine the equilibrium E-pH lines for the couples Co2+/Co and CoO/Co, as shown in Fig. 5. The two lines intersect at point P, and above them are the domains of stability for Co 2+ and CoO. The boundary between these two = 1 (Fig. 5). domains is a vertical line containing point P and located at pH 6.3 for

Fig. 5 Partial E-pH diagram for the Co2+/Co and CoO/Co couples for

=1

Figure 6(a) shows a partial E-pH diagram for different values of in which only three chemical 2+ species—Co, Co , and CoO (or the hydroxide Co(OH)2)—are considered. There are, however, other possible

chemical species, dissolved or solid, that must be considered. This introduces new equilibria that modify the diagram to give Fig. 6(b).

Fig. 6 E-pH diagram for the cobalt-water system for various values of log diagram. (b) Complete E-pH diagram

. (a) Partial E-pH

The Water E-pH Diagram. Pourbaix diagrams are traced for equilibrium reactions taking place in water; consequently, the water system must always be considered at the same time as the system under investigation in any E-pH diagram. Water can be decomposed into oxygen and hydrogen, according to the following electrode reactions: (Eq 20) and

Considering the equilibrium of water dissociation/ionization into solvated protons and hydroxide ions: equilibria (Eq 20) may also be written as: and

The equilibrium potentials of these two electrochemical reactions can be determined by using the Nernst equation. For the water/hydrogen or proton/hydrogen couple:

(Eq 21)

where

is the fugacity or pressure of hydrogen near the electrode (in fact, dimensionless fugacity is equal to

the numerical value of the fugacity expressed in bar). At 25 °C (298.15 K), because, by definition,

=0

V/SHE (see the article “Electrode Potentials” in this Section of the Volume), the previous equation can be rewritten as: (Eq 22) This relation for

Fig. 6(b) and 7 by line a, which decreases with increasing pH.

Fig. 7 The water E-pH diagram at 25 °C (298.15 K) and 1 bar For the oxygen/water or oxygen/hydroxide couple:

(Eq 23)

is the fugacity of O2 near the electrode. The water activity is, as usual, assumed to be 1. At 25 °C where (298.15 K), the standard potential for O2/H2O is 1.23 V SHE (Ref 4), so the equation can be rewritten as: (Eq 24) This relation for is represented by line b in Fig. 6(b) and 7. It is interesting to note that the pressures of hydrogen and oxygen in the vicinity of the electrode are usually identical and nearly equal to the pressure that exists in the electrochemical cell. To be rigorous, the water vapor pressure should be taken into account, but it is frequently neglected as not being very significant at 25 °C (298.15 K). When the pressure increases, line b in Fig. 7 is displaced upward in the diagram, and line a is lowered. The result is that the domain of water stability increases with increasing pressure. The water system is very important for a good understanding of the corrosion behavior of metals; it is represented (usually by dashed lines) in all Pourbaix diagrams (Ref 4). Conventions for E-pH Diagram Construction. In the construction of diagrams for binary (metal-water) systems, the authors follow the original Pourbaix format (Ref 4) and delineate regions within which condensed phases are stable by solid lines, whereas coexistence lines separating the predominance areas for dissolved species are drawn dashed, even in the regions where the condensed species are stable (Fig. 8). Also, it is common to draw the lines of separation between solid compounds and dissolved species for a number of different activities of the latter. The large number of lines may render the diagram difficult to read (Fig. 8). Moreover, the amount of the chemical element considered, when summed over all dissolved species containing this element, should be constant over the diagram. On diagrams calculated on this principle of constant total element concentration, as the activities in the ideal solution model are taken equal to the numerical values of the concentrations

(molalities), the lines corresponding to the limits of the stability domain of a solid species are rounded where two solution species coexist (e.g., the boundaries between domains of Fe2+, Fe3+, and Fe2O3 in Fig. 8).

Fig. 8 Original Pourbaix diagram for the iron-water system at 25 °C (298.15 K) (oxides are considered; hydroxides are not). Source: Ref 4 When this principle is adopted rigorously, the diagrams are tedious to compute. Therefore widely adopted convention in calculating the E- pH line for the equilibrium between a solid species and a dissolved species is that the dissolved species has simply an activity equal to the molality equal to that it would have if it were the only form present, that is, the selected molality of the element in solution, divided by the number of atoms of the element in a molecule of the species. The effect of this simplifying assumption is that all lines on the diagrams are straight. Figure 9 shows a diagram simplified with respect to the original Pourbaix format (Fig. 8). The diagram (Fig. 9) is further simplified by limiting the pH range from 0 to 14.

Fig. 9 Simplified E-pH diagram for the iron-water system at 25 °C for a molality of dissolved iron equal to 10-6 mol/kg. Pressure of hydrogen and oxygen, 1 atm Consider the consequences of choosing a convention on the element concentration when constructing a diagram. For an equilibrium between two dissolved forms, R and O, of an element. The E-pH line calculated for equal concentration of the two forms separates the domains of relative predominance of R and O is calculated from the Nernst equation for equal concentrations of the element under the two forms. When the chemical formulae of the two forms contain the same number of atoms of the element, the concentrations of the two forms are equal, so the logarithmic term with the ratio of activities in the Nernst equation is equal to 0. In this case, the equation of the separation line does not depend on the total concentration of the element (Ref 4, 6, 41). For example, for the reaction: (Eq 25) the equilibrium equation is:

(Eq 26)

The equation of the limit between the areas of relative predominance of

and H2S(aq) simplifies to: (Eq 27)

This is not true when the formula of the two dissolved forms do not contain the same number of atoms of the element. For example, for the reaction: (Eq 28) the equilibrium equation is:

(Eq 29)

In this case, the equation of the separation line depends on the total concentration of sulfur (molality mS). If it were considered that the total element amount is constant, the equation of the limit of the areas of relative predominance would correspond to the case where half of sulfur is as H2S(aq) and the other half as S2

.

This would be obtained by taking the concentration mS/2 for H2S(aq) and mS/4 for S2 equation of the separation line would then be:

(Ref 4, 6, 41). The

or (Eq 30) If the simplifying convention described previously is chosen, the E-pH line for equilibrium between two dissolved species is calculated considering that each dissolved species is the only form of the element present in solution; for example, taking the concentration mS for H2S(aq) and mS/2 for S2 separation line is:

. Then, the equation of the

or

(Eq 31)

The simplification with respect to the classic Pourbaix presentation of the diagrams allows a gain of clarity and makes easier the construction of the more complex diagrams for multicomponent systems. Diagrams for Metastable Species. It is to be noted that the species present in solution in real conditions are not necessarily the more stable ones but may be metastable species that are less stable thermodynamically but, for kinetic reasons, are the ones effectively present in solution for a finite time. A classic example is the sulfurwater system where the oxidized forms of sulfur present in solution are not necessarily the thermodynamically stable species elemental solid sulfur, hydrogenosulfate,

, and sulfate,

metastable dithionates S2

) sulfites (H2SO3,

, hydrosulfites (HS2

, S2

, ions but can be the , and

),

tetrathionates (S4 ), or thiosulfates (HS2 , S2 ) (Ref 4, 6, 7, 41). Possible reactions between the metal and sulfur metastable species are of interest for prediction of corrosion in industrial systems. The E-pH diagram for the sulfur-water system showing the thermodynamically stable species is given in Fig. 10 for a sulfur activity (molarity) of 10-4 mol/kg. It exhibits a small stability domain of solid sulfur from acid to neutral pHs. The same diagram shows the E-pH lines for the case where the thiosulfate species are considered as the only oxidized forms of sulfur (dashed lines). According to the construction conventions, the activity of the dissolved sulfides containing one sulfur atom per molecule or ion is taken as 10-4 and the activity of thiosulfate species containing two sulfur atoms per ion as 0.5 × 10-4.

Fig. 10 Simplified E-pH diagram for the sulfur-water system at 25 °C. The solid lines represent the stable system. The dashed lines represent the equilibria involving the metastable thiosulfates instead of the stable sulfates. Sulfur molality is 10-4 mol/kg. Pressure of hydrogen and oxygen, 1 atm

References cited in this section 4. M. Pourbaix, Atlas of the Electrochemical Equilibria, NACE, 1974 5. D.D. Wagman, W.H. Evans, V.B. Parker, R.H. Schumm, I. Halow, S.M. Bailey, K.L. Churney, and R.L. Nuttall, J. Phys. Chem. Ref. Data, Vol 11 (Suppl. 2), 1982 6. G. Valensi, “Rapport CEBELCOR-CEFA/ R. 17,” 1958; “Rapport Technique CEBELCOR, 121, RT. 207, 1,” 1973 7. R.C. Murray, Jr. and D. Cubicciotti, J. Electrochem. Soc., Vol 130, 1983, p 866 41. M. Pourbaix and A. Pourbaix, Rapports Techniques CEBELCOR, 159, RT. 299 (1990)


Practical Use of E-pH Diagrams The E-pH diagram is an important tool for understanding electrochemical phenomena. It provides useful thermodynamic information in a simple figure. Two cases are presented here to illustrate its practical use in corrosion prediction. Corrosion of Nickel. A rod of nickel is immersed in an aqueous deaerated acid solution that contains 10-4 mol/L of Ni2+ ions. The system is at 25 °C (77 °F) under 1 atm pressure. The E-pH diagram corresponding to these conditions is shown in Fig. 11.

Fig. 11 E-pH diagram for nickel for

= 10-4

At the metal-water interface, two electrochemical reactions are possible: Ni

Ni2+ + 2e-

(Eq 32)

and 2H+ + 2e-

(Eq 33)

H2

The equilibrium potentials of the Ni2+/Ni and the H+/H2 electrodes can be computed. From Eq 6, −0.25 + 0.030 log

, which, for

-4

= 10 , gives

≈ −0.37 V (Fig. 2). From Eq 22,

= =

−0.059 pH, which, at pH = 1, for example, gives ≈ −0.06 V. up to a pH of approximately 6 (Fig. 11). Thus, when connected via an electrical circuit, electrons tend to flow from the more negative nickel electrode, where they are produced by the oxidation reaction, to the less negative hydrogen electrode, where they are consumed by the reduction reaction. In this case, the two electrodes are formed on the surface of the conducting nickel rod. The two reactions will proceed under a common electrode potential or mixed potential, with a value somewhere between the nickel and hydrogen equilibrium potentials. Up to pH ≈ 6, the mixed potential, EM, is located above the Ni2+/Ni equilibrium potential in the region of Ni2+ stability and below the H+/H2 equilibrium potential in the region of H2 stability (Fig. 11). Hence, nickel is not stable at low pH in water, and it tends to oxidize (corrode) into Ni2+, while H+ is reduced into hydrogen gas (hydrogen evolution).

Thus, the Pourbaix diagram explains the tendency for nickel to corrode in acid solutions. It does not indicate the rate of corrosion, however. This important information has to be obtained from a kinetic experiment, for example, from the recording of current versus potential curve around the corrosion potential. The Pourbaix diagram also shows that when the pH increases to approximately 6, the difference between the nickel and the hydrogen equilibrium potential decreases in magnitude, and consequently, the corrosion tendency diminishes. For pHs between 6 and 8 (limit of stability of Ni 2+), Fig. 11 shows that the hydrogen electrode potential becomes lower than the nickel one. Under these conditions, H+ can no longer accept the electrons from nickel. The mixed potential of the system is, in this case, below the equilibrium potential of Ni2+/Ni, in the region of metal immunity. Hence, in water at room temperature, nickel does not corrode for pH 6 to 8. No such pH range exists for the iron-water system, where the Fe2+/Fe electrode potential is always lower than the H+/H2 electrode potential (Fig. 1, 8, 9). Therefore, iron will always corrode to ferrous ions with evolution of H2 in acid and neutral solutions. In contrast, the behavior of nickel makes this metal slightly noble (in the small pH range of 6 to 8), and, from the diagram, it is expected to resist corrosion better than iron. Moreover, an increase in hydrogen pressure, according to Eq 22, lowers the equilibrium line of H+/H2 while it does not change the equilibrium line of Ni2+/Ni. As a result, an increase in pressure leads to greater corrosion resistance for nickel. For pHs higher than 8, films of NiO (or the hydrated form Ni(OH)2) and Ni3O4 can form at the surface at anodic potentials, as can be seen in Fig. 11. These oxides may, in some cases, protect the metal by forming a protective layer that prevents or mitigates further corrosion. This phenomenon is called passivation. It also occurs on iron with the formation of magnetite (Fe3O4) or hematite (Fe2O3) at anodic potentials (Fig. 8, 9). The presence of species such as chlorine ions may increase the corrosion tendency of metals, because these species may attack the protective layer and then favor corrosion. Figures 1, 8, and 11 illustrate that iron or nickel may corrode (dissolve) in very strong alkaline solutions as or , respectively, or, more likely, as the hydrated forms or . Corrosion of Copper. Observation of the copper E-pH diagram in Fig. 12 immediately reveals that the corrosion of copper immersed in deaerated acid water is not likely to occur. The H+/H2 equilibrium potential represented by line a is always lower than the Cu2+/Cu equilibrium potential. The H+ ions are stable in contact with copper metal, which cannot corrode (is immune) in water solutions free from oxidizing agents.

Fig. 12 Partial E-pH diagram for copper for

= 10-4

The presence of dissolved oxygen in nondeaerated solutions introduces another possible reaction: O2 reduction into H2O, with an equilibrium potential higher than that of Cu2+/Cu. The O2/H2O couple is then a good acceptor for the electrons produced by copper oxidation. The two electrochemical reactions: Cu → Cu2+ + 2eand

(Eq 34)

(Eq 35)

O2 + 2H+ + 2e- → H2O

take place spontaneously in acid solutions at the surface of an immersed piece of copper at a common (mixed) potential. In neutral or alkaline solutions, O2 reduction will be coupled with copper oxidation into Cu2O.


E-pH Diagrams for Ternary Systems The E-pH diagrams for corrosion protection give valuable information if all the substances present in the actual system under investigation (metal or metalloid in aqueous solution) are taken into account when the diagrams are constructed. The previous discussion assumed that the solution did not contain chlorine, sulfur, or other species capable of forming solid compounds or soluble complexes with the metal. In the presence of such elements, other diagrams must be considered that may reveal different metal corrosion behavior. Hence, diagrams for binary metal-water (M-H2O) systems are of limited usefulness, and diagrams for multicomponent systems must be calculated. For example, the simple diagram of gold in water (Au-H2O system) does not show any solubility for that metal. The addition of cyanide (CN) ions to the system, however, leads to the formation of a gold complex soluble in water. Hence, gold, which does not corrode in pure water can dissolve in the presence of cyanide. This property is the basis of gold plating and of the hydrometallurgy of that metal. Diagrams for the ternary system Au- CN--H2O must then be constructed to show the conditions of stability for the gold complex. A general bibliography of E-pH diagrams for multicomponent systems in aqueous solutions is given in Ref 8. Diagrams for ternary metal-additive-water (M-A-H2O) systems are frequently used. In most cases, the amount of one element in the dissolved form is substantially greater than that of the other. If so, the reactions involving only the major element will be practically independent of the reactions involving the minor element. Thus, the diagram of the major element alone, drawn with dashed lines, will serve as a background for the diagram for the combined system, drawn with solid lines. The behavior of the minor element may be or may not be independent of that of the major element. It will be independent if the minor element forms no stable solid compound or dissolved complex with the major element; in this case the diagram for the ternary system will just be the diagram of the minor element alone superimposed on the diagram of the major element alone. Otherwise the diagram for the ternary system may be drastically different. Again, the convention is that in the regions of stability of a dissolved species, the latter is considered as the only element form present. When the minor element forms a dissolved complex with the major element, this solution species is considered as a species of the minor element, and hence its concentration is the one of the minor element. Consider the case where the concentration of additive in the dissolved form is much higher than the tolerated concentration of dissolved metal. First, the E-pH diagram for the A-H2O system must be calculated to determine the areas of predominance for the various species of A. Then, in each area delineated with dashed lines, the E-pH equilibria lines for the various reactions between the metal and the predominant A species to form metal compounds or complexes are calculated. A metal-additive dissolved complex is considered as a metal species. As an example of calculations for a ternary system involving metal-additive solid compounds, consider the following reactions for the iron-sulfur-water system: •

FeS2/Fe2+ equilibrium in sulfide media:

FeS2 + 4H+ + 2e-

•

or

Fe2+ + 2H2S(aq)

(Eq 36)

FeS2 + 2H+ + 2e-

•

Fe2+ + 2HS-

(Eq 37)

Fe2+/FeS2 equilibrium in thiosulfate media:

(Eq 38)

•

or

(Eq 39)

•

Fe2O3/FeS2 equilibrium in thiosulfate media:

(Eq 40)

At 25 °C (77 °F), the equations for the previously mentioned equilibria are, respectively, if mS and mFe are the molalities of sulfur and iron in solution:

(Eq 41)

or

(Eq 42)

(Eq 43)

or

(Eq 44)

(Eq 45)

Metals may form compounds in dissolved form with sulfur (e.g., FeSO4(aq), 2+

Cu /

). Consider the

equilibrium in sulfite media in a case where mCu « mS:

(Eq 46)

(Eq 47)

The equation of the boundary line between the domains of relative predominance of Cu2+ and (= mCu): obtained (if the latter were stable) for

E = Eo + 0.059 log mS

would be

(Eq 48)

An illustration is given for the system iron- sulfur-water in water containing thiosulfates. This case is of technological importance, because thiosulfates dissolved in aqueous solution are known to be detrimental to the corrosion resistance of stainless steels (Ref 9, 10, 11, 12, 13, 14, 15, 16). The concentration of the dissolved sulfur impurity expressed in molality is usually approximately 10-4 mol/kg, and the molality of dissolved iron is considered to be 10-6 mol/kg. Such a small value allows conservative predictions of corrosion, because it is generally agreed that there is no corrosion when the concentration of metal that can be dissolved in a solution initially free from it is lower or equal to 10-6 mol/kg. Thus, the major element here is sulfur, and the binary diagram used as a background is the one of the metastable sulfur-water system (compare with Fig. 10), showing the thermodynamically stable sulfides (H2S(aq) and HS-) as reduced forms of sulfur and the metastable thiosulfates ( iron-sulfur-water system is plotted in Fig. 13.

and

) as the only oxidized forms. The ternary diagram for the

Fig. 13 E-pH diagram for the iron-sulfur-water system at 25 °C (298.15 K) in the case where the metastable thiosulfates are the only oxidized forms of sulfur. The stability domains are limited by the dotted lines for the water system, dashed lines for the sulfur-water system, and solid lines for the iron-sulfur-water system. mS = 10-4 mol/kg. mFe = 10-6 mol/kg

A comparison with the diagram of the binary iron-water system (Fig. 8, 9) shows that iron interacts with sulfides or thiosulfates in the mid- pH region to form iron sufides (FeS and FeS2), replacing Fe2+ in acid solutions and magnetite (Fe3O4) in neutral and alkaline solutions. Because metal sulfides are good ionic conductors, they offer little protection against corrosion. Although the diagram predicts Fe2O3 could be formed on the surface under anodic (strongly oxidizing) conditions over a large pH range and protect iron from corrosion, the incompatibility between FeS2 and Fe2O3 actually prevents growth of an adhesive Fe2O3 layer (Ref 17). Hence, the diagram predicts that iron passivation will not occur in the presence of sulfides or thiosulfates in solution. The diagram for the chromium-sulfur-water system is plotted in Fig. 14. There is no range of stability of chromium sulfides (at least for mS ≤ 10-4 mol/kg), so the diagram is identical to the binary diagram for the chromium-water system. It shows that chromium sulfides are less stable than the chromic oxide (Cr2O3) or hydroxide (Cr(OH)3), which provide the exceptional corrosion resistance of chromium.

Fig. 14 E-pH diagram for the chromium-sulfur-water system at 25 °C (298.15 K) in the case where the thiosulfates are the only oxidized forms of sulfur. The stability domains are limited by the dotted lines for the water system, dashed lines for the sulfur-water system, and solid lines for the chromium-sulfur-water system. mS = 10-4 mol/kg and mCr = 10-6 mol/kg

References cited in this section 8. J.V. Muylder, in Comprehensive Treatise of Electrochemistry, Vol 4, J. O'M. Bockris, B.E. Conway, E. Yeager, and R.E. White, Ed., Plenum Press, 1981, p 1–96 9. R.C. Newman, H.S. Isaacs, and B. Alman, Corrosion, Vol 38, 1982, p 261

10. R.C. Newman K. Sieradski, and H.S. Isaacs, Metall. Trans. A, Vol 13, 1982, p 2015 11. R.C. Newman and K. Sieradski, Corros. Sci., Vol 23, 1983, p 363 12. R.C. Newman, Corrosion, Vol 41, 1985, p 450 13. D. Tromans and L. Frederick, Corrosion, Vol 40, 1984, p 633 14. A. Garner, Corrosion, Vol 41, 1985, p 587 15. S.E. Lott and R.C. Alkire, J. Electrochem. Soc., Vol 136, 1989, p 973, 3256 16. C. Duret-Thual, D. Costa, W.P. Yang, and P. Marcus, Corros. Sci., Vol 39, 1997, p 913 17. C.M. Chen, K. Aral, and G.J. Theus, “Computer-Calculated Potential pH Diagrams to 300 °C,” EPRI Report NP3137, Vol 2, Electric Power Research Institute, Babcock & Wilcox Company, June 1983


E-pH Diagrams for High- Temperature Aqueous Solutions Prediction of the corrosion behavior of metals in aqueous solutions at high temperatures is of considerable technological interest. Such a situation is steel vessels in contact with pressurized water at 300 °C (573.15 K) in nuclear power reactors. It was necessary to extend the E-pH equilibrium diagrams originally established at 25 °C (Ref 4) to higher temperatures. The main problem is that there are few thermodynamic data available for chemical species dissolved in water above 60 °C (333.15 K). The increase in water vapor pressure with temperature in closed cells leads to high pressures. This makes it necessary to carry out experiments in autoclaves, high- strength vessels with special seals (Ref 18). Often, thermodynamic-state functions data at high temperatures are obtained by extrapolation from the data at 25 °C using empirical hypotheses on the influence of temperature on molar heat capacities. In particular, entropy correspondence principles have been established that make possible the calculation of entropies and heat capacities of ions at temperatures up to 300 °C from the entropy values at 25 °C (Ref 19, 20). An alternate treatment of the problem is the calculation of temperature coefficients of standard equilibrium potentials of electrochemical reactions from thermodynamic data at 25 °C (Ref 21). A review of these different methods of estimation is given in Ref 8. High-temperature Pourbaix diagrams are, by nature, less accurate than those at 25 °C that are based on experimental data, because they are based on estimates. The effects of pressure on the equilibria can be ignored up to 300 °C because the magnitude of the errors introduced is within the uncertainty of the data (Ref 18, 19). At high temperature, the molar scale for activities must not be used, because, due to water volume changes, the molarity for a given solute quantity (number of moles of solute per liter of water) varies with temperature. One can only use the molality (number of moles of solute per kilogram of water), which is equal to the molarity at 25 °C (77 °F) but is invariant with temperature. The standard state for each dissolved species is the hypothetical ideal solution of the substance at unit molality. Thermodynamic Conventions for High Temperature. High-temperature thermodynamic calculations for aqueous solutions require one to specify additional conventions, and one must be careful about which convention applies when using a set of high temperature data.

Standard chemical potentials, or Gibbs free energies of formation at a temperature T, may be given for the formation of the substance from its elements in their standard states at T or at 25 °C (77 °F). Thus, in the first convention, the standard values of elements or diatomic gases are zero at any temperature, whereas, in the second one, they are zero only at 25 °C (77 °F). Using the second convention, values are given in Table 1 for hydrospecies and sulfur species at 25 and 300 °C. Table 2 lists the values for iron and its compounds. Table 1 Standard Gibbs free energies of formation (chemical potentials) of hydrospecies and sulfur compounds (sulfides and thiosulfates only) at 25 and 300 °C Species H2(g) O2(g) S(rh,l) H2S(g) H2O(l) H+(aq) e-(aq) H2S(aq) HSHS2 S2

ΔfG0(298.15 K), kJ/mol 0 0 0 -33.282 -237.174 0 0 -27.861 12.050 -532.414 -522.582

ΔfG0(573.15 K), kJ/mol -38.786 -59.400 -12.678 -93.423 -263.881 0 -19.393 -78.881 13.648 -564.840 -505.260

References 21 21 21 21 22 … … 22 7 22 7 7

Table 2 Standard Gibbs free energies of formation (chemical potentials) of iron compounds at 25 and 300 °C Species Fe Fe3O4 Fe2O3 Fe2+ FeOH+ Fe3+ FeOH2+

ΔfG0(298.15 K), kJ/mol 0 -1015.457 -742.242 -91.563 -275.542 -614.989 -17.280 -229.409 -438.065

ΔfG0(573.15 K), kJ/mol -10.155 -1072.468 -777.906 -57.497 -272.542 -629.094(a) 62.718 -193.548(a) -438.270(a)

References 17 17 17 22 22 17 22 17 17

-467.290

-375.836(a)

17

-100.754 -124.540 23 FeS -160.168 -181.440 23 FeS2 (a) Values after correction of the difference in convention for the free energy of formation of H+(aq) at 300 °C (573.15 K) between this text and Ref 17 Two conventions exist for the potential scale: the “universal” convention and the “alternate” convention. In the universal convention, the electrode potentials are referred to the potential of the standard hydrogen electrode is taken equal (SHE) at the temperature considered; that is, the potential of the SHE to 0 V at all temperatures. Hence, the chemical potential (Gibbs free energy) of the conventional electron used in the writing of electrode (half- cell) reactions is, at all temperatures: (Eq 49) With this convention, the E-pH line for the H+(aq)/H2(g) couple passes by the point (E = 0, pH = 0) at any T (Fig. 13). In the alternate convention, the potentials are referred to the potential of the SHE at 25 °C (77 °F). Here, the chemical potential of the conventional electron is the same at all temperatures:

(Eq 50) In this convention, the potential of the SHE depends on temperature, and the E-pH line for the H+(aq)/H2(g) couple intersects the vertical axis for pH = 0 above 0 V for T > 25 °C (Fig. 15). The two conventions become identical at 25 °C:

(298.15 K) =

= 0.

Fig. 15 pH at 100 and 300 °C (373.15 and 573.15 K) versus pH at 25 °C (298.15 K) in an unbuffered solution (pH200 = pH300). Also, the correspondence between the pH scales at the different temperatures is shown. Solvated H+ (H+(aq)) is often considered as a reference substance whose standard chemical potential, or Gibbs free energy of formation, is taken to be zero at all temperatures. Use of the universal convention produces the = . simple relation Variation of pH with Temperature. For a correct interpretation of the high-temperature E- pH diagrams, the change of pH of the solution with the temperature must be taken into account. The pH of an aqueous solution is determined by: •

•

The equilibrium constant for water dissociation into solvated protons and hydroxide ions, H2O(l) = H+(aq) + OH-(aq), also called the ionic product of water, (Kw)T = , which increases with temperature The concentrations and dissociation constants of the other constituents of the solution

Because the temperature dependence of dissociation constants is not the same for all acids and bases, it is not possible to calculate the pH scale for a given temperature in a manner that would apply to all aqueous solutions. At least, the temperature effect can be visualized on the E-pH diagram at a given temperature by marking by vertical lines the position of three important pH values (Ref 24): • •

pH = -log10(Kw)T = 0, which corresponds to a 1 molal solution of a strong (completely dissociated) acid pH = -log10(Kw)T = (pKw)T, which corresponds to a 1 molal solution of a strong base

•

pH = (pKw)T, which corresponds to neutral water. Indeed, in neutral aqueous solutions:

Hence, the neutral pH is pHn = = − (log10(Kw)T) = (pKw)T. The pKw decreases from 14.00 at 25 °C to 12.27 at 100 °C (373.15 K) and 11.30 at 200 and 300 (473.15 and 573.15 K) (Ref 22). Accordingly, the pH of a neutral aqueous solution is 7.00 at 25 °C, whereas it is 6.13 at 100 °C and 5.65 at 200 and 300 °C. In an unbuffered solution (with completely dissociated acid or base), the proton activity is fixed only by the water ionic product (from here simply denoted KT), which increases with T. Thus, a solution that has a certain pH at 25 °C will have a lower pH at a higher temperature. In this simple case, the change in pH can be calculated in the following way (Ref 23). Consider the general case of a solution of a certain pH at 25 °C: From the definition of the water ionic product K25 = , hence pK25 = pH25 + pOH25. As the temperature is raised, the equilibrium of dissociation of water will be shifted, and equivalent amounts of additional H+ and OH- will be generated in solution. Thus: (Eq 51) where x is the increase in ion activity as a result of the temperature change from 25 °C to T. This is a quadratic equation in x: x2 +

+ (K25 - KT) = 0, whose physically meaningful solution is:

(Eq 52)

The pH at T is pHT = •

•

•

. It takes limiting values:

The pH values at 100 and 300 °C are plotted versus the pH at 25 °C 15 in Fig. 15. Also, the correspondence between the pH scales at different temperatures is plotted. (The pH at 200 °C is equal to the pH at 300 °C, because pKw is practically the same at these two temperatures [Ref 22].) To compare the behavior of an electrode in an unbuffered solution of given pH at 25 °C with the one in the same solution at T, the corrected pH must be employed when using the E-pH diagram at T. It must be noted that, in a buffered solution, the pH is fixed by the equilibrium of an acid- base couple, and it is the variation of the acid dissociation constant with temperature that is predominant in fixing the pH variation. Temperature Effects on the E-pH Diagrams for Binary Systems. The diagram for the iron- water system at 300 °C is plotted in Fig. 16. Even after correction of the temperature effect on pH, the net effect of increase of temperature is to shift the diagram to lower values of pH (Ref 17, 23, 24, 25). Comparison of highertemperature diagrams to the diagram at 25 °C (Fig. 8) gives the following trends for the influence of temperature: •

•

The domains of stability of the Fe2+ and Fe3+ cations are contracted to the benefit of condensed species Fe, Fe3O4, and Fe2O3; that is, the solubility of the latter species in acid solutions is lower at 300 °C than at 25 °C. There is an expansion below pH 14 of the domain of stability of the dihypoferrite ion (or ) at the expense of Fe, Fe3O4, and Fe2O3; that is, the solubility of the condensed species in alkaline solutions is substantially greater at 300 °C.

Fig. 16 E-pH diagram for the iron-water system at 300 °C (573.15 K). mFe = 10-6 mol/kg The diagram for the chromium-water system at 300 °C (573.15 K) is plotted in Fig. 17. Compared to the diagram at 25 °C (Fig. 14), CrOH2+ replaces Cr3+ as the trivalent species at low pH and anodic potentials. Effects of temperature similar to those described for the iron-water system are observed. The Cr2O3 (or Cr(OH)3) becomes less soluble (more stable) in acid solution and more soluble (less stable) in alkaline solutions, where

is the stable ion (Ref 17, 26, 27, 28).

Fig. 17 E-pH diagram for the chromium-water system at 300 °C (573.15 K). mCr = 10-6 mol/kg Similar effects are also predicted for various other metals (Ref 17, 18, 26, 27). Thus, the most significant effect of the increase of temperature is an expansion of the domain of corrosion in strong alkaline environment, which was confirmed experimentally (Ref 8, 23). Temperature Effects on E-pH Diagrams for Ternary Systems. Diagrams for multi-component systems in hightemperature aqueous solutions are obviously of great interest for predicting corrosion behavior in numerous industrial conditions. A general bibliography of E-pH diagrams for multicomponent systems in hightemperature aqueous solutions is given in Ref 8. As an example, the diagram for the ternary system iron-sulfurwater with thiosulfates at 300 °C (573.15 K), for aS = 10-4 and aFe = 10-6, is shown in Fig. 18. It may be compared to the diagram at 25 °C, presented previously (Fig. 13), to visualize the effects of increasing temperature. Concerning the sulfur-water system, there is an increased stability of the acid forms of dissolved , at the expense of HS- and S2 . At 300 °C, the diagram for the iron-sulfur-water sulfur, H2S(aq) and HS2 system is identical to the one for the iron-water system (Fig. 16). Besides the effects described previously for the iron-water system (contraction of the domains of stability of Fe2+ and Fe3+ and expansion of the domain of ), the main effect of temperature rise is that the range of stability of the sulphides FeS and FeS2 is drastically reduced (Ref 17, 29) and even suppressed for the sulfur activity of the diagram (aS = 10-4).

Fig. 18 E-pH diagram for the iron-sulfur-water system at 300 °C (573.15 K) in the case where the thiosulfates are the only oxidized forms of sulfur. mS = 10-4 mol/kg, mFe = 10-6 mol/kg Similarly, the diagram for the chromium-sulfur-water system at 300 °C (572 °F) is identical to the diagram for the chromium-water system shown in Fig. 17, because it shows no chromium sulfides.

High-temperature E-pH diagrams for various metal-chlorine-water systems, which are very important for the interpretation and prediction of corrosion phenomena in high-salinity brines, can be found in Ref 30. Also, the E-pH diagrams for the quaternary system Fe-Cl-S-H2O up to 250 °C (523.15 K) are of direct interest in the phenomenon of stress cracking in sulfide-containing brines (Ref 31).

References cited in this section 4. M. Pourbaix, Atlas of the Electrochemical Equilibria, NACE, 1974 7. R.C. Murray, Jr. and D. Cubicciotti, J. Electrochem. Soc., Vol 130, 1983, p 866 8. J.V. Muylder, in Comprehensive Treatise of Electrochemistry, Vol 4, J. O'M. Bockris, B.E. Conway, E. Yeager, and R.E. White, Ed., Plenum Press, 1981, p 1–96 17. C.M. Chen, K. Aral, and G.J. Theus, “Computer-Calculated Potential pH Diagrams to 300 °C,” EPRI Report NP-3137, Vol 2, Electric Power Research Institute, Babcock & Wilcox Company, June 1983 18. R.L. Cowan and R.W. Staehle, J. Electrochem. Soc., Vol 118, 1971, p 557 19. C.M. Criss and J.W. Cobble, J. Am. Chem. Soc., Vol 86, 1964, p 5385, 5390, 5394 20. I.L. Khodakovski, B.N. Ryzhenko, and G.B. Naumov, Geochem. Int., Vol 5, 1968, p 1200, translated from Geokhimia, No. 12, 1968, p 1486; I.L. Khodakovski, Geochem. Int., Vol 6, 1969, p 29, translated from Geokhimia, No. 1, 1969, p 57 21. A.J. de Béthune, T.S. Licht, and N. Swendeman, J. Electrochem. Soc., Vol 106, 1959, p 616; G.R. Salvi and A.J. de Béthune, J. Electrochem. Soc., Vol 108, 1961, p 672 22. J.W. Cobble, R.C. Murray, Jr., P.J. Turner, and K. Chen, “High-Temperature Thermodynamic Data for Species in Aqueous Solution,” San Diego State University Foundation, EPRI Report NP-2400, Electric Power Research Institute, May 1982 23. V. Ashworth and P.J. Boden, Corros. Sci., Vol 10, 1970, p 709 24. H.E. Townsend, in Proceedings of the Fourth International Congress on Metallic Corrosion, 1969 (Amsterdam), NACE, 1970, p 477 25. H.E. Townsend, Corros. Sci., Vol 10, 1970, p 343 26. P.A. Brook, Corros. Sci., Vol 12, 1972, p 297 27. J.B. Lee, Corrosion, Vol 37, 1981, p 467 28. P. Radhakrishnamurty and P. Adaikkalam, Corros. Sci., Vol 22, 1982, p 753 29. R.J. Biernat and R.G. Robins, Electrochim. Acta, Vol 17, 1972, p 1261 30. D.D. MacDonald, B.C. Syrett, and S.S. Wing, Corrosion, Vol 35, 1979, p 1 31. D.D. MacDonald, “ASTM Sp. Publication 717,” ASTM, 1981


E-pH Diagrams for Adsorbed Species The principle of E-pH diagrams can be extended to the case of bidimensional layers of species adsorbed on metal surfaces. The solid compounds treated in the usual diagrams are three-dimensional (bulk) compounds (oxides, hydroxides, sulfides, etc.). However, the formation of a three-dimensional solid compound M xAy (A may be O, OH, S …) by reaction of gaseous or dissolved forms of A with a metal, M, is often preceded by the formation of a two- dimensional phase of A adsorbed on the metal surface. This surface phase is more stable than the bulk compound (Ref 32). When there is creation of a true chemical bond between A atoms and metal surface atoms (bond energies larger than 200 kJ/mol, at least at low coverage), the adsorption is also called chemisorption. Adsorbed (chemisorbed) monolayers may form under E-pH conditions where the bulk compounds are thermodynamically unstable, and a classic diagram would predict only the existence of the bare metal. Chemisorption must not be neglected, because the presence of a chemisorbed monolayer can induce marked changes in the reactivity of the metal. For example, it has been shown that a monolayer of sulfur adsorbed on nickel or nickel-iron alloys enhances the anodic dissolution and hinders the formation of the passive film, drastically affecting the corrosion resistance of the metallic material (Ref 33, 34). Therefore, E-pH diagrams for adsorbed species are of interest for predicting corrosion risk (Ref 34). The method of calculation of the equilibrium potentials of oxidation-reduction couples involving a species adsorbed on an electrode surface is presented subsequently. Principle of E-pH Diagrams for Adsorbed Species. Consider the case where a species A (which may be an element, H, O, S or a molecular species, OH or H2O) is adsorbed from solution on a metal surface M in the form of a neutral monolayer [denoted Aads(M)]. The adsorption of A from a dissolved species in aqueous solution may result from an electrooxidation or an electroreduction reaction, depending on the valence state of A in the dissolved species (it may then be called electroadsorption). The adsorption of an atom or molecule on a metal surface in water involves the replacement of adsorbed water molecules and a competition with the adsorption of oxygen species (atomic oxygen O or hydroxyl OH) produced by electrooxidation of water or hydroxide (OH-) ions present in the electrolyte. For simplicity, a Langmuir model for adsorption is taken in which it is assumed that the two-dimensional (surface) phase is an ideal substitutional solution where species adsorb competitively on the same sites without lateral interactions between adsorbed species. Under these conditions, the chemical potential of each adsorbed species Aads(M) in the surface phase of a metal M can be expressed as follows: (Eq 53) where θA is the relative coverage by adsorbed A (0 ≤ θA ≤ 1; θA = 1 for a complete monolayer of A); is the standard chemical potential of A, corresponding to the saturation of the surface by Aads(M). The total coverage on the surface is equal to unity; the coverage by water is = (1 − ∑θAi), where θAi is the coverage by any other adsorbed species. As an example of the method of calculation of the E-pH relations, consider the adsorption of hydroxyl on a metal surface from adsorbed water: H2Oads(M)

OHads(M) + H+ + e-

(Eq 54)

The equilibrium potential of this half-cell reaction is obtained by applying the Nernst equation, using the expression of the chemical potentials of adsorbed hydroxyl and water, as given in Eq 53:

(Eq 55)

where θOH and are the relative coverages by adsorbed hydroxyl and water on the M surface. The standard potential Eo on the SHE scale is given by: (Eq 56) The standard chemical potentials, or Gibbs free energies of formation, for adsorbed sulfur oxygen, or hydroxyl are derived from literature thermodynamic data on the reversible chemisorption of these species at the metalwater interface or, if not available, at the metal-gas interface (Ref 35, 36, 37, 38, 39). Adsorbed water is weakly adsorbed on most transition metals, compared to chemisorbed species. In the absence of accurate thermodynamic data, the standard Gibbs free energy of adsorption of water from the liquid state may be neglected; that is, the standard Gibbs free energy of adsorbed water is approximated by the free energy of liquid water. This value and the values for sulfur and oxygen adsorbed on iron, calculated for the formation of the chemisorbed species from S and O2 in their standard state at 25 °C, are listed in Table 3. Table 3 Standard Gibbs free energies of formation (chemical potentials) for water, oxygen, and sulfur adsorbed on iron surfaces at 25 and 300 °C Species ΔfG0(298.15 K), kJ/mol ΔfG0(573.15 K), kJ/mol References -263.881(a) 36 H2Oads(Fe) -237.174(a) -229 -236 36 Oads(Fe) -176 -188 36 Sads(Fe) (a) The values taken for H2Oads(Fe) are the values for H2O(l). Source: Ref 22 E-pH Equations for Oxygen and Sulfur Adsorbed on Iron. As an example, presented here are the E-pH relations for the equilibria between oxygen and sulfur adsorbed on iron and dissolved species in water containing ). It is considered that the ratio of the area of sulfides (H2S or HS-) or thiosulfates (HS2 electrode to the volume of solution is small enough so that the activity (molality mS) of each dissolved sulfur species is independent of the sulfur surface coverage, θS. The uncertainty on these equations is relatively high because of the lack of accuracy of the thermodynamic measurements presently available for chemisorbed sulfur and oxygen: Sads(Fe) + H2O(l)

Oads(Fe) + H2S(aq)

(Eq 57)

At 25 °C: log(θO/θS) = -18.8 - log mS

(Eq 58)

At 300 °C: log(θO/θS) = -7.9 - log mS

(Eq 59)

where θO and θS are the relative coverages by the adsorbed oxygen and sulfur on the Fe surface. Sads(Fe) + H2O(l)

Oads(Fe) + HS- + H+

(Eq 60)

At 25 °C: log(θO/θS) = pH - 25.8 - log mS

(Eq 61)

At 300 °C: log(θO/θS) = pH - 16.4 - log mS H2Oads(Fe)

Oads(Fe) + 2H+ + 2e-

(Eq 62) (Eq 63)

(Eq 64)

(Eq 65)

(Eq 66)

(Eq 67)

(Eq 68)

(Eq 69)

(Eq 70)

(Eq 71)

(Eq 72)

(Eq 73)

(Eq 74)

(Eq 75)

(Eq 76)

(Eq 77) Potential-pH Diagrams for Oxygen and Sulfur Adsorbed on Iron. The preceding equations have been used to plot the E-pH diagrams for sulfur and oxygen adsorbed on iron in water containing sulfides or thiosulfates (Ref

40). The diagrams at 25 and 300 °C for a molality of dissolved sulfur mS = 10-4 mol/kg are shown in Fig. 19 and 20. The diagrams are superimposed on the iron-sulfur-water diagrams described previously (Fig. 13, 18).

Fig. 19 E-pH diagram for the system of sulfur, oxygen, and water adsorbed on iron at 25 °C (298.15 K) in the case where the thiosulfates are the only oxidized forms of sulfur. The stability domains are limited by the dotted lines for the water system, dashed lines for the sulfur-water system, and thin solid lines for the iron-sulfur-water system and thick solid lines for the adsorbed species system. θS and θO are the relative surface coverages of adsorbed sulfur and oxygen, respectively. mS = 10-4 mol/kg, mFe = 10-6 mol/kg

Fig. 20 E-pH diagram for the system of sulfur, oxygen, and water adsorbed on iron at 300 °C (573.15 K) in the case where the thiosulfates are the only oxidized forms of sulfur. The stability domains are limited by the dotted lines for the water system, dashed lines for the sulfur-water system, and thin solid lines for the iron-sulfur-water system and thick solid lines for the adsorbed species system. θS and θO are the relative surface coverages of adsorbed sulfur and oxygen, respectively. mS = 10-4 mol/kg, mFe = 10-6 mol/kg The domains of stability of adsorbed species are limited by lines corresponding to significant values of the surface coverage: θ = 0.01; 0.5; 0.99. For θS > 0.5 and a ratio θO/θS < 0.01, sulfur is considered as the only adsorbed species in the domain, and θO may be neglected in Eq 51, Eq 52, Eq 52, Eq 52, Eq 52. Similarly, for θO > 0.5 and a ratio θO/θS > 100, oxygen is considered as the only adsorbed species, and θS may be neglected in Eq 64 and 65. For 0.01 < θO/θS < 100, the adsorbed phase is a mixture of coadsorbed sulfur and oxygen, and both terms θS and θO must be taken into account in the equations. In the stability domain of H2S(aq), the ratio θO/θS is constant (Eq 58, 59). It is negligible at 25 and 300 °C (for mS = 10-4 mol/kg), hence, θO can be neglected in Eq 67 and 68, and only sulfur is adsorbed by replacement of ( = 1 − θS - θO) is fixed, and the E-pH relation for water. For a given value of θS, the ratio sulfur adsorption from H2S(aq) by replacement of water (Eq 67, 68) gives a straight line (Fig. 19, 20). In the stability domain of HS-, the ratio θO/θS increases with pH, according to Eq 61 and 62. At 25 °C, this ratio is infinitesimal up to pH 14. At 300 °C for mS = 10-4 mol/kg, it becomes significant (>0.01) for a pH value below 14. The pH values corresponding to θO/θS = 0.01 (pH = 10.4) and θO/θS = 1 (pH = 12.4) are represented by vertical lines in the diagram (Fig. 20). The first vertical line is the left boundary of a domain, where sulfur and oxygen are coadsorbed. In this domain, as the ratio θO/θS varies with pH, the ratios θS/(1 - θS - θO) and θO(1 - θS - θO) depend both on θS (or θO) and pH. Therefore, the E-pH relations calculated for sulfur adsorption from HS- (Eq 70, 71) and oxygen adsorption from water (Eq 64, 65), for given values of θS and θO, give nonstraight

lines (Fig. 20). These lines become vertical at the pH values where the water coverage becomes infinitesimal, that is, the coverage θS + θO reaches unity (full monolayer of coadsorbed sulfur and oxygen). At these pHs, the line for a given sulfur coverage, θS, meets the line for the complementary oxygen coverage θO = 1 - θS, and they merge with the vertical line plotted for the corresponding ratio θO/θS (Fig. 20). In the stability domains of thiosulfates, the E- pH relations for the replacement reaction between adsorbed sulfur and oxygen give straight lines for a fixed ratio θO/θS (Eq 73, 74, 76, 77). A simplification occurs here, because the water coverage = 1 − θO - θS is negligible in the domain of thiosulfates (that can be checked by associating Eq 58 and 59 or 61 and 62 with 64 and 65 and calculating at the anodic limit of the sulfides domains), so θO in Eq 73, 74, 76, and 77 can be approximated by 1 - θS. Then, straight lines are obtained for given values of θS, which delimit the respective stability domains of adsorbed sulfur and oxygen (Fig. 19, 20). The diagrams (Fig. 19, 20) allow the prediction of the E-pH conditions in which sulfur is adsorbed on an iron surface from sulfides or from thiosulfates dissolved in water (Ref 40). The main features are the following: when the potential is increased, adsorbed water molecules are replaced by sulfur atoms adsorbed by electrooxidation of sulfides. Similarly, when the potential is decreased, adsorbed oxygen atoms (or hydroxyl groups) are replaced, totally or partially, by sulfur atoms adsorbed by electroreduction of thiosulfates. The replacement takes place within a very narrow range of potential (~0.06 V at 25 °C; ~0.11 V at 300 °C). At 300 °C, the stability domain of adsorbed sulfur alone is limited at high pH by the domain of coadsorption Sads - Oads (Ref 40). The two-dimensional reactions involving oxygen (hydroxyl) and sulfur adsorbed on bare iron surfaces are of a different nature than the reactions involving the three-dimensional (bulk) Fe- O(OH) or iron-sulfur compounds. Hence, the diagrams developed here (Fig. 19, 20) are different from the classic E-pH diagrams of the ironsulfur-water system (Fig. 13, 18). However, superimposition of the two types of diagrams is useful to discuss in more detail the possible effects of an adsorbed sulfur layer on the corrosion behavior of iron. At room temperature, the domain of stability of the adsorbed sulfur monolayer includes the stability domains of the bulk metal sulfides and Fe3O4 and overlaps the domains of metallic iron (immunity domain) of Fe2+ (activity domain), and Fe2O3 (Fig. 19). Sulfur adsorption is then expected for E-pH conditions where iron sulfides are not thermodynamically stable, which reflects the excess of stability of the two- dimensional chemisorbed species with respect to the three-dimensional compounds. At 300 °C, whereas no region of stability of iron sulfide exists for mS = 10-4 mol/kg, adsorbed sulfur is stable in a large domain, which includes the domain of Fe3O4 and overlaps the domains of iron, Fe2+, Fe2O3, and (Fig. 20). The prediction of domains of thermodynamic stability of adsorbed sulfur on iron in thiosulfate solutions supports the experimental observation of sulfur adsorption by thiosulfate reduction on iron-chromium alloys, which was invoked to explain the detrimental effect of dissolved thiosulfates on the corrosion resistance of ferritic stainless steels (Ref 16). The diagrams indicate stability of Sads in a large part of the passivity domain. The chemisorption of sulfur on bare iron is a process in competition with the formation of Fe3O4 (and Fe2O3 in a limited E-pH region). The equilibrium E-pH diagrams are constructed on a thermodynamic basis and do not indicate which species actually form on a bare iron electrode polarized in the passive domain: the twodimensional (surface) species Sads(Fe) or a three-dimensional (bulk) oxide. If the kinetics of adsorption of sulfur on bare iron is more rapid than the kinetics of formation of oxide layers on iron, a sulfur monolayer may form on iron and prevent or delay passivation of the iron. Detrimental effects of sulfur on the corrosion resistance of iron are then expected, even under E- pH conditions where a classic diagram predicts passivity. The diagrams (Fig. 19, 20) also predict that sulfur is likely to adsorb in part of the domain of anodic dissolution of iron; this is important because dissolution enhanced by adsorbed sulfur is experimentally observed in the activity domains of nickel- and iron-base alloys (Ref 33, 34). Even if the metal is not thermodynamically stable and dissolves, a sulfur monolayer may adsorb on the fresh surface, which is continuously produced, and increase the kinetics of dissolution. Thus, the E-pH diagrams presented here showing the domains of thermodynamic stability of adsorbed layers on metals, provide a basis for assessing the risk of corrosion of metals or alloys induced by species adsorbed from aqueous solutions.

References cited in this section

16. C. Duret-Thual, D. Costa, W.P. Yang, and P. Marcus, Corros. Sci., Vol 39, 1997, p 913 22. J.W. Cobble, R.C. Murray, Jr., P.J. Turner, and K. Chen, “High-Temperature Thermodynamic Data for Species in Aqueous Solution,” San Diego State University Foundation, EPRI Report NP-2400, Electric Power Research Institute, May 1982 32. J. Oudar, in Corrosion Mechanisms in Theory and Practice, P. Marcus, Ed., Marcel Dekker, Inc., 2002 33. J. Oudar and P. Marcus, Appl. Surf. Sci., Vol 3, 1979, p 48 34. P. Marcus, in Corrosion Mechanisms in Theory and Practice, P. Marcus, Ed., Marcel Dekker, Inc., 2002, p 287 35. P. Marcus and E. Protopopoff, C.R. Acad. Sci. Paris, Vol 308 (No. II), 1989, p 1685 36. P. Marcus and E. Protopopoff, J. Electrochem. Soc., Vol 137, 1990, p 2709 37. P. Marcus and E. Protopopoff, J. Electrochem. Soc., Vol 140, 1993, 1571 38. P. Marcus and E. Protopopoff, J. Electrochem. Soc., Vol 144, 1997, p 1586 39. P. Marcus and E. Protopopoff, Corros. Sci., Vol 45, 2003, p 1191 40. P. Marcus and E. Protopopoff, Corros. Sci., Vol 39, 1997, p 1741–1752


Acknowledgment Portions of this article have been adapted from D.L. Piron, Potential versus pH (Pourbaix) Diagrams, Corrosion, Vol 13, Metals Handbook, 9th ed., ASM International, 1987, p 24–28.


References

1. M. Pourbaix, Thermodynamique des Solutions Aqueuses Diluées, Potentiel D'oxydo- Réduction (résumé de conférence), Bull. Soc. Chim. Belgique, Vol 48, Dec 1938 2. M. Pourbaix, Thermodynamics of Dilute Aqueous Solutions, Arnold Publications, 1949 3. R.W. Staehle, Marcel J.N. Pourbaix—Palladium Award Medalist, J. Eletrochem. Soc., Vol 123, 1976, p 23C 4. M. Pourbaix, Atlas of the Electrochemical Equilibria, NACE, 1974 5. D.D. Wagman, W.H. Evans, V.B. Parker, R.H. Schumm, I. Halow, S.M. Bailey, K.L. Churney, and R.L. Nuttall, J. Phys. Chem. Ref. Data, Vol 11 (Suppl. 2), 1982 6. G. Valensi, “Rapport CEBELCOR-CEFA/ R. 17,” 1958; “Rapport Technique CEBELCOR, 121, RT. 207, 1,” 1973 7. R.C. Murray, Jr. and D. Cubicciotti, J. Electrochem. Soc., Vol 130, 1983, p 866 8. J.V. Muylder, in Comprehensive Treatise of Electrochemistry, Vol 4, J. O'M. Bockris, B.E. Conway, E. Yeager, and R.E. White, Ed., Plenum Press, 1981, p 1–96 9. R.C. Newman, H.S. Isaacs, and B. Alman, Corrosion, Vol 38, 1982, p 261 10. R.C. Newman K. Sieradski, and H.S. Isaacs, Metall. Trans. A, Vol 13, 1982, p 2015 11. R.C. Newman and K. Sieradski, Corros. Sci., Vol 23, 1983, p 363 12. R.C. Newman, Corrosion, Vol 41, 1985, p 450 13. D. Tromans and L. Frederick, Corrosion, Vol 40, 1984, p 633 14. A. Garner, Corrosion, Vol 41, 1985, p 587 15. S.E. Lott and R.C. Alkire, J. Electrochem. Soc., Vol 136, 1989, p 973, 3256 16. C. Duret-Thual, D. Costa, W.P. Yang, and P. Marcus, Corros. Sci., Vol 39, 1997, p 913 17. C.M. Chen, K. Aral, and G.J. Theus, “Computer-Calculated Potential pH Diagrams to 300 °C,” EPRI Report NP-3137, Vol 2, Electric Power Research Institute, Babcock & Wilcox Company, June 1983 18. R.L. Cowan and R.W. Staehle, J. Electrochem. Soc., Vol 118, 1971, p 557 19. C.M. Criss and J.W. Cobble, J. Am. Chem. Soc., Vol 86, 1964, p 5385, 5390, 5394 20. I.L. Khodakovski, B.N. Ryzhenko, and G.B. Naumov, Geochem. Int., Vol 5, 1968, p 1200, translated from Geokhimia, No. 12, 1968, p 1486; I.L. Khodakovski, Geochem. Int., Vol 6, 1969, p 29, translated from Geokhimia, No. 1, 1969, p 57 21. A.J. de Béthune, T.S. Licht, and N. Swendeman, J. Electrochem. Soc., Vol 106, 1959, p 616; G.R. Salvi and A.J. de Béthune, J. Electrochem. Soc., Vol 108, 1961, p 672 22. J.W. Cobble, R.C. Murray, Jr., P.J. Turner, and K. Chen, “High-Temperature Thermodynamic Data for Species in Aqueous Solution,” San Diego State University Foundation, EPRI Report NP-2400, Electric Power Research Institute, May 1982

23. V. Ashworth and P.J. Boden, Corros. Sci., Vol 10, 1970, p 709 24. H.E. Townsend, in Proceedings of the Fourth International Congress on Metallic Corrosion, 1969 (Amsterdam), NACE, 1970, p 477 25. H.E. Townsend, Corros. Sci., Vol 10, 1970, p 343 26. P.A. Brook, Corros. Sci., Vol 12, 1972, p 297 27. J.B. Lee, Corrosion, Vol 37, 1981, p 467 28. P. Radhakrishnamurty and P. Adaikkalam, Corros. Sci., Vol 22, 1982, p 753 29. R.J. Biernat and R.G. Robins, Electrochim. Acta, Vol 17, 1972, p 1261 30. D.D. MacDonald, B.C. Syrett, and S.S. Wing, Corrosion, Vol 35, 1979, p 1 31. D.D. MacDonald, “ASTM Sp. Publication 717,” ASTM, 1981 32. J. Oudar, in Corrosion Mechanisms in Theory and Practice, P. Marcus, Ed., Marcel Dekker, Inc., 2002 33. J. Oudar and P. Marcus, Appl. Surf. Sci., Vol 3, 1979, p 48 34. P. Marcus, in Corrosion Mechanisms in Theory and Practice, P. Marcus, Ed., Marcel Dekker, Inc., 2002, p 287 35. P. Marcus and E. Protopopoff, C.R. Acad. Sci. Paris, Vol 308 (No. II), 1989, p 1685 36. P. Marcus and E. Protopopoff, J. Electrochem. Soc., Vol 137, 1990, p 2709 37. P. Marcus and E. Protopopoff, J. Electrochem. Soc., Vol 140, 1993, 1571 38. P. Marcus and E. Protopopoff, J. Electrochem. Soc., Vol 144, 1997, p 1586 39. P. Marcus and E. Protopopoff, Corros. Sci., Vol 45, 2003, p 1191 40. P. Marcus and E. Protopopoff, Corros. Sci., Vol 39, 1997, p 1741–1752 41. M. Pourbaix and A. Pourbaix, Rapports Techniques CEBELCOR, 159, RT. 299 (1990)


Selected References

• •

C.M. Chen, K. Aral, and G.J. Theus, “Computer Calculated Potential pH Diagrams to 300 °C,” Vol 1, 2, and 3, EPRI NP-3137, Project 1167-2, Electric Power Research Institute, June 1983 R.P. Frankenthal and J. Kruger, Ed., Equilibrium Diagrams/Localized Corrosion, Proceedings of an International Symposium to Honor Marcel Pourbaix on His Eightieth Birthday, Vol 84–89, The Electrochemical Society, 1984, p 611

K.H. Stern, Molten Salt Corrosion Thermodynamics, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 31–33

Molten Salt Corrosion Thermodynamics Kurt H. Stern, Naval Research Laboratory

Introduction MOLTEN SALTS—in contrast to aqueous solutions in which an electrolyte (acid, base, salt) is dissolved in a molecular solvent—are essentially completely ionic. Thus the terms solute and solvent can be defined only in quantitative terms. For example, the terms lose their meaning in a NaCl-AgCl melt where composition can vary continuously from pure NaCl to pure AgCl. This is true even when the electrodes immersed in the melt are reversible only to some of the ions in the melt. For example, in the cell: (Eq 1)

Ag|AgCl|1NaCl|Cl2 -

the chlorine electrode is reversible to Cl and the silver electrode is reversible to Ag+. When this cell is used to obtain thermodynamic data, it is assumed that the cell is stable; that is, its composition does not change with time. However, when the concentration of AgCl is very low, this will not be the case, since the silver electrode will react spontaneously with the melt: Ag + NaCl = AgCl + Na

(Eq 2)

Thus the concentration of AgCl will spontaneously increase in the melt, and the electromotive forces (emf) measured for the preceding cell will not be stable but will change in the direction indicating increasing Ag concentration. The point at which this happens depends on the system.

K.H. Stern, Molten Salt Corrosion Thermodynamics, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 31–33 Molten Salt Corrosion Thermodynamics Kurt H. Stern, Naval Research Laboratory

Thermodynamics of Cells One major use of electrochemical cells is to obtain thermodynamic data for salts. The basic thermodynamics applicable to galvanic cells for aqueous solutions is discussed elsewhere in this Volume. Only those aspects that are different for molten salts are emphasized in this article. Thus, for the cell given in Eq 1, the cell reactions are:

Ag = Ag+ + e-

(Eq 3)

Cl2 + e- = Cl-

and the result is: (Eq 4)

Ag + Cl2 = AgCl(a) where a is the activity of AgCl and the cell emf is

(Eq 5) It is assumed that when Cl2 is in its standard state of unit fugacity (nearly equal to unit pressure) and silver is in its standard state (pure metal), Eo is the cell emf for: Ag(a = 1) + Cl2(f = 1) = AgCl(a = 1)

(Eq 6)

and the only variable in Eq 5 is the activity of AgCl. (It should be emphasized that in any thermodynamic treatment, only neutral components; that is, AgCl, rather than Ag + or Cl-, can appear.) In contrast to aqueous solutions, the activity, a, is usually defined on the mole fraction scale. The emf is E for the cell reaction shown previously, where: ΔG = -nFE

(Eq 7)

and (Eq 8) The superscript o(ΔGo, Eo) refers to the standard state, which in this case is the pure material; that is, AgCl. An activity coefficient, γ, is defined in terms of the activity, a, and the mole fraction, X: γ = a/X Note that γ may be less, equal, or greater than unity. This definition is based on Raoult's law standard state: γ = 1 when a = X

(Eq 9)

In the case of an electrochemical cell, using Eq 1 as an example, where the cell emf is directly related to the activity of AgCl, the activity of the second component, NaCl, can then be calculated by the Gibbs-Duhem equation, which at constant temperature and pressure is: ∑ Xi d ln ai = 0

(Eq 10)

For the AgCl-NaCl system, the activity of AgCl(a1) is known from experimental data (Eq 5), and the activity of NaCl(a2) can then be calculated from Eq 10. The procedure for doing this is discussed in thermodynamic texts such as Ref 1. It generally requires a graphical integration as discussed in standard thermodynamics texts as well as standard free energies of formation (ΔGo) of the pure components, which can be obtained from reference works like Ref 2. This equation shows that if the composition varies, the chemical potentials of the components do not change independently but instead in a related way. In contrast to aqueous solutions where the potential of the hydrogen electrode ( H2 = H+ + e-) is commonly assigned the value “zero” when reactants and products are in their standard states, no such condition exists for molten salts, because there is no solvent analogous to water in molten salts. Thus it is necessary to establish a specific scale for each medium; for example, Cl2 + e- = Cl-. It should be noted that while metal electrodes, Ag in the previous example, can be directly immersed in the melt, gas electrodes, like the chlorine electrode, require a solid surface on which the equilibrium Cl2 + e- = Cl- can take place. Graphite has frequently been used for chlorine; platinum is often chosen for other gases, such as oxygen.

References cited in this section 1. G.N. Lewis and M. Randall, revised by K.S. Pitzer and D.F. Brewer, Thermodynamics, 2nd ed., McGraw-Hill, 1961, p 552

2. G.J. Janz, Molten Salt Handbook, Academic Press, 1967


Electrodes for Molten Salts The classification of electrodes is similar to that used for aqueous media. There are reference electrodes that maintain a constant potential independent of the melt concentration, and indicator electrodes that are reversible to some ion in the melt and can therefore be used to measure its activity. Reference Electrodes. Historically, the oldest reference electrodes are those that use a liquid junction (LJ) between the electrode compartment and the melt under study. The Ag/AgCl electrode is of this type, and its half-cell is: (Eq 11) The melt under study contains the same ions as the reference melt, but its concentration is variable. If the concentrations of the two melts are not too far apart, the LJ potential is probably negligible. One way of forming the junction is to insert an asbestos fiber into a small hole between the compartments. It may also be formed inside a material that is porous and thus can be saturated with the melt but that prevents bulk mixing. Jenkins, Mamantov, and Manning (Ref 3) described a reference for use in molten fluorides based on the Ni/Ni(II) couple, which uses boron nitride as the separator. The boron nitride becomes saturated by the melt and thus serves as an LJ that isolates the reference melt. Porous glass can also be used as a separator, but obviously not in fluoride melts. Another reference electrode useful in melts that contain oxide ion, usually as Na2O, was developed by Liang, Bowen, and Elliott (Ref 4). This electrode is: (Eq 12) and operates at an oxide activity fixed by the reaction: W(s) + Na2S + O2 = WS(s) + Na2O

(Eq 13)

inside a beta alumina membrane, where S is solid. Alternatively, mullite, a ceramic that is also a sodium ion conductor, can also be used as a separator that isolates the reference electrode from the melt under study. This is particularly useful because mullite closed-end tubes are readily available, and the reference electrode and its associated melt are then simply inserted into the tube. For melts containing dissolved oxides, the oxygen electrode inside a zirconia membrane is a useful reference. The actual electrode is platinum bathed in an environment of O2 at a fixed partial pressure. Air is usually used as the source of oxygen. The platinum is forced into a zirconia tube stabilized with a divalent oxide, such as CaO, MgO, or Y2O3, which acts as an oxide ion conductor. The electrode reaction is: O2 + 2e = O2-

(Eq 14)

Indicator Electrodes. The chief characteristic of indicator electrodes is that their potential varies with the activity of an ionic component, usually in a Nernstian way. The performance of molten salt indicator electrodes does not differ substantially from the performance of other electrodes. For example, the Ag electrode in the melt under study changes its potential with changes in melt composition (Eq 5). However, because of the

technical problems in maintaining an LJ, such as corrosion, most indicator electrodes use solid-state materials that are ionic conductors: stabilized zirconia, which is an oxide-ion conductor; beta-alumina, a sodium-ion conductor; and sulfide-ion conductors, such as Cu2S. For an example of how the silver reference electrode can be combined with other electrodes to measure oxide ion activity, see Ref 5. In selecting these materials, the range of temperatures and gas phase composition over which the material is an ionic conductor must be considered. If a substantial fraction of the conductivity is electronic, the response of the electrode will not be Nernstian, and emf measurements will give inaccurate results. For each material there is a diagram that separates the ionic conductivity from the electronic (n-type and p-type) as a function of pressure and temperature (Ref 6), and it is obviously desirable that most of the conductivity is ionic if the material is to be useful. Materials based on stabilized zirconia are particularly useful, because they are oxide ion conductors. The electrical properties of these materials have been described in detail (Ref 7). The uses of these materials fall into two categories: equilibrium methods and nonequilibrium methods. Examples of equilibrium methods include: • • • •

Measurement of the O2 partial pressure in gases (Ref 7) Measurement of the O2 solubility in metals (Ref 7) Measurement of the dissociation pressure of solid oxides (Ref 7) Measurement of the oxide activity in molten salts (Ref 8, 9, 10) and in glasses (Ref 11)

Nonequilibrium methods include: • • • •

Operation of an oxygen fuel cell (Ref 4) Changing the O2 content of gases (Ref 12, 13) Measurement of the diffusivity of O2 in metals (Ref 14) Titration of O2 into and out of metals (Ref 14) and molten salts (Ref 15)

References cited in this section 3. H.W. Jenkins, G. Mamantov, and D.L. Manning, J. Electroanal. Chem., Vol 19, 1968, p 385 4. W. Liang, H.K. Bowen, and J.F. Elliott, in Metal-Slag-Gas Reactions and Processes, A.Z. Foroulis and W.W. Smeltzer, Ed., Electrochemical Society, 1975, p 608 5. K.H. Stern, M.L. Deanhardt, and Rm. Panayappan, J. Phys. Chem., Vol 83, 1979, p 2848 6. W.L. Worrell and J. Hladik, Chapter 17, Vol 1 in Physics of Electrolytes, J. Hladik, Ed., Academic Press, 1972 7. T.H. Etsell and S.N. Flengas, Chem. Revs., Vol 70, 1970, p 339 8. R. Combes, J. Vedel, and B. Tremillon, Anal. Lett., Vol 3, 1970, p 523 9. D.R. Flinn and K.H. Stern, J. Electroanal. Chem., Vol 63, 1975, p 39 10. K.H. Stern, Rm. Panayappan, and D.R. Flinn, J. Electrochem. Soc., Vol 124, 1977, p 641 11. G.C. Charette and S.N. Flengas, Can. Metall., Vol 7, 1969, p 191 12. C. Deportes, R. Donneau, and G. Robert, Bull. Soc. Chim. France, 1964, p 2221 13. D. Yuan and F.A. Kroger, J. Electrochem. Soc., Vol 116, 1969, p 594 14. R.L. Pastorek and R.A. Rapp, Trans. Met. Soc. AIME, Vol 25, 1969, p 1711

15. M.L. Deanhardt and K.H. Stern, J. Phys. Chem., Vol 84, 1980, p 2831


Thermodynamics of Molten Salt Corrosion Corrosion of metals in aqueous solution has been studied for many years, and the thermodynamic principles have been discussed by Pourbaix (Ref 16, 17). The applicable principles are commonly summarized in Pourbaix diagrams, which contain a great deal of data in a convenient arrangement that allows the prediction of what reactions will occur in a particular system. A corresponding system for molten salt corrosion has been explicitly described for molten chlorides (Ref 18, 19). In contrast to aqueous solutions where corrosion is always accompanied by oxidation, corrosion in molten salts can be caused by the solubility of the metal in the salt, particularly if the metal dissolves in its own chloride. The more usual case is one in which the metal is oxidized. Littlewood has used the convention that the free energy of formation of the halide ion from the halogen gas at one atmosphere pressure and unit halide activity is zero at all temperatures. Gibbs energies (ΔG) are converted to potentials by the usual relation E = -ΔG/nF. In Pourbaix diagrams, equilibrium potentials are plotted against pH with a resulting schematic division into regions of stability of different solid phases. For molten salts, Littlewood chose the activity of oxide in the melt, expressed as pO2-(p = -log), and the appropriate diagram is then E versus pO2-. Such diagrams can be constructed for any metal and melt, and Littlewood has described several cases (Ref 20), including those in which there is more than one stable oxide, such as titanium. These diagrams should be as useful for molten salts as Pourbaix diagrams are for aqueous solutions. However, although this work was published about 30 years ago, it does not seem to have been widely used. Such a diagram that plots potential as a function of pO2- is given in Fig. 1 for the Ti-NaCl system at 800 °C and Fig. 2 for the Ti-MgCl2 system. Note that O2- is the concentration on the mole fraction scale and pO2- is its negative logarithm. On both diagrams, the outer scale, labeled log O2, refers to the pressure of oxygen gas in each system.

Fig. 1 E-pO2- diagram for Ti-NaCl system at 800 °C (approximate). Source: Ref 20

Fig. 2 E-pO2- diagram for Ti-MgCl2 system at 800 °C (approximate). Source: Ref 20 The corrosion of titanium is affected by three factors: the titanium chloride content of the MgCl2 melt, Mg or Na content, and oxygen content of the product. The interplay of the three factors can be readily investigated by a detailed comparison of Fig. 1 and 2. Comparison of the diagrams shows that for a given value of pO2-, the oxygen activity in the titanium metal product would be much lower for magnesium reduction than for sodium reduction. The equilibrium potentials after the reaction has gone to completion will be -3.20V for sodium and 2.47V for magnesium. Suppose, for example, the oxide contamination present in the two cases is sufficient to produce a pO2- value of 20 in the slag and that excess reductant has been used. Figure 1 shows that under these

conditions, the oxygen activity in the titanium metal produced by sodium reduction would be about 10-62, while that produced by magnesium reduction (Fig. 2) would be about 10-86.

References cited in this section 16. M. Pourbaix, Thermodynamics of Dilute Aqueous Solutions, Arnold, London, 1949 17. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, NACE International, 1974 18. C. Edeleanu and R. Littlewood, Electrochim. Acta, Vol 3, 1960, p 195 19. R. Littlewood and E.J. Argent, Electrochim. Acta, Vol 4, 1961, p114, 155 20. R. Littlewood, J. Electrochem. Soc., Vol 109, 1962, p 525


References 1. G.N. Lewis and M. Randall, revised by K.S. Pitzer and D.F. Brewer, Thermodynamics, 2nd ed., McGraw-Hill, 1961, p 552 2. G.J. Janz, Molten Salt Handbook, Academic Press, 1967 3. H.W. Jenkins, G. Mamantov, and D.L. Manning, J. Electroanal. Chem., Vol 19, 1968, p 385 4. W. Liang, H.K. Bowen, and J.F. Elliott, in Metal-Slag-Gas Reactions and Processes, A.Z. Foroulis and W.W. Smeltzer, Ed., Electrochemical Society, 1975, p 608 5. K.H. Stern, M.L. Deanhardt, and Rm. Panayappan, J. Phys. Chem., Vol 83, 1979, p 2848 6. W.L. Worrell and J. Hladik, Chapter 17, Vol 1 in Physics of Electrolytes, J. Hladik, Ed., Academic Press, 1972 7. T.H. Etsell and S.N. Flengas, Chem. Revs., Vol 70, 1970, p 339 8. R. Combes, J. Vedel, and B. Tremillon, Anal. Lett., Vol 3, 1970, p 523 9. D.R. Flinn and K.H. Stern, J. Electroanal. Chem., Vol 63, 1975, p 39 10. K.H. Stern, Rm. Panayappan, and D.R. Flinn, J. Electrochem. Soc., Vol 124, 1977, p 641 11. G.C. Charette and S.N. Flengas, Can. Metall., Vol 7, 1969, p 191 12. C. Deportes, R. Donneau, and G. Robert, Bull. Soc. Chim. France, 1964, p 2221

13. D. Yuan and F.A. Kroger, J. Electrochem. Soc., Vol 116, 1969, p 594 14. R.L. Pastorek and R.A. Rapp, Trans. Met. Soc. AIME, Vol 25, 1969, p 1711 15. M.L. Deanhardt and K.H. Stern, J. Phys. Chem., Vol 84, 1980, p 2831 16. M. Pourbaix, Thermodynamics of Dilute Aqueous Solutions, Arnold, London, 1949 17. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, NACE International, 1974 18. C. Edeleanu and R. Littlewood, Electrochim. Acta, Vol 3, 1960, p 195 19. R. Littlewood and E.J. Argent, Electrochim. Acta, Vol 4, 1961, p114, 155 20. R. Littlewood, J. Electrochem. Soc., Vol 109, 1962, p 525


Selected References • •

J. Hladik, Ed., Physics of Electrolytes, Vol 1; Transport Properties in Solid Electrolytes, Vol 2, Thermodynamics and Electrode Processes in Solid State Electrolytes, Academic Press, 1972 R.W. Laity, D.J.G. Ives, and G. J. Janz, Ed., Reference Electrodes, Chapter 12, Electrodes in Fused Salt Systems, Academic Press, 1961

S.A. Matthes, Geochemical Modeling, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 34–41

Geochemical Modeling Steven A. Matthes, U.S. Department of Energy, Albany Research Center

Introduction CORROSION SCIENTISTS AND ENGINEERS frequently need to describe the chemical state of local environments, including mass distribution of aqueous species, and the interaction of these aqueous species with metal surfaces, corrosion films, and gases. In high temperature corrosion, metals may be corroded by nonaqueous fluids or by solids such as slags. These types of interactions are geochemical in nature; that is, the reactions are the same as those found in natural aqueous, high-temperature, or magmatic systems. When

chemical systems are simple with few possible reactions likely to occur, the geochemistry of a local environment may be described through experience and some (usually tedious) hand calculation. When the geochemistry of an environment is more complex, however, quantitative models of aqueous or solid phase chemistry that are beyond the predictive power of experience and hand calculation are required. The need for a consistent way to evaluate complex geochemical systems has led to a rapid increase in the field of geochemical modeling. The rapid growth roughly parallels the rise in power and reduction in cost of computers needed to evaluate the complicated quantitative models of solution chemistry. Geochemical modeling is now being used to understand and predict scaling, susceptibility to corrosion, atmospheric corrosion rates, acid rain, corrosion film solubility, and environmental impacts of aqueous species in runoff. Geochemical modeling software (GMS) is widely used by scientists and engineers to solve a variety of materials and energy-balance problems. The predictive ability of GMS allows the user to model real situations and to ascertain how materials should behave when subjected to different environments. The advent of inexpensive and powerful desktop computers and a wide variety of both commercial and freeware software tools make this predictive power available to almost every scientist and engineer. As an example, the freeware program PHREEQC is capable of simulating a wide range of aqueous geochemical reactions including the mixing of waters, addition of net irreversible reactions to solution, dissolving and precipitating phases to achieve equilibrium with the aqueous phase, and the effects of changing temperature. The PHREEQC program and commercial software packages such as AquaChem (Fig. 1) now have sophisticated graphical user interfaces for performing geochemical modeling.

Fig. 1 Determination of equilibrium composition using the geochemical modeling software (GMS) AquaChem Concentrations of elements, molarities and activities of aqueous species, pH, saturation indices, and the electrochemical oxidation state of the system are calculated with this type of software. Mole transfers of phases to achieve equilibrium can also be calculated as a function of specific reversible and irreversible reactions, provided the necessary thermodynamic data are available.

The electrochemical oxidation state of aqueous systems is usually given as the negative log of the electron activity (pe) or the electrochemical oxidation potential (EH) in geochemical modeling. Just as pH is useful in describing the equilibrium position of all acid/base pairs in a system:

pe expresses the equilibrium position of all redox pairs in a system. For example, for the Fe+3/ Fe+2 redox couple,

Just as pH may be considered the controlling variable for the acid/base system, pe is the controlling variable for the redox system. Treating the electron like other reactants and products in a system allows redox reactions such as the previous one to be combined directly with other reactions leading to the determination of equilibrium constants for the combined reactions of interest in a system. Although EH is usually thought of as the measured potential of an aqueous system, it is also related to the pe by the equation: EH = (2.303RTk/F)pe where R is the gas constant, F is the Faraday constant, and Tk is the absolute temperature. The environments that GMS best describes, aqueous systems that are open or closed to the atmosphere and that have interactions with one or more solid phases, are precisely the same environments that concern corrosion engineers. Software systems such as The Geochemist's Workbench can easily model aqueous systems in contact with several mineral phases (Fig. 2). Ironically, GMS systems are seldom used by corrosion scientists and engineers, even though geochemical modeling may provide greater insight into the causes and results of corrosion reactions and processes.

Fig. 2 A plot of log of oxygen activity (log a O2 (aqueous) versus pH) for arsenic minerals using The Geochemist's Workbench. Log a O2 is log of oxidation potential.

S.A. Matthes, Geochemical Modeling, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 34–41 Geochemical Modeling Steven A. Matthes, U.S. Department of Energy, Albany Research Center

History of Geochemical Modeling In 1947, an algorithm for the numerical solution of equilibrium states in a multicomponent system was published (Ref 1). Although intended for a desk calculator, the method was quickly adopted for use with digital computers. The algorithm was based on the evaluation of equations for equilibrium constants. Garrels and Thompson (Ref 2) calculated (by hand) the composition of seawater in terms of dissociated ions (Na+, Ca++, , etc.). This calculation gave rise to a type of geochemical model that predicts species distributions, saturation indices, and gas fugacities from chemical analyses. Equilibrium models of this type have become widely used, largely because of the availability of software such as SOLMNEQ (Ref 3), WATEQ (Ref 4), and MINEQL (Ref 5). Another type of geochemical model, the reaction path model, was introduced in 1967 (Ref 6) and was used to model the effects of evaporation (gradually increasing concentration) on species distribution. This model extended geochemical modeling from the evaluation of a fixed system to having the ability to simulate a process. Hegelson (Ref 7) built on previous work by coding the general equations for species distribution and mass transfer into a computer program. The program Path1 (Ref 8) was the first general computer method for modeling reaction paths such as weathering, evaporation, ore deposition, and hydrothermal alteration. One way of modeling reaction paths is by repeatedly solving for the system equilibrium state algebraically. Karpov and Kaz'min (Ref 9) and Karpov et al. (Ref 10) applied this idea to GMS, allowing an algebraic solution to systems varying in temperature and composition. In 1979, EQ3/EQ6 (Ref 11) was introduced as the first geochemical model to be documented and distributed based on algebraic equations. This greatly simplified the computer code and separated mass and heat transfer calculations from the chemical equilibrium calculations. Parkhurst et al. introduced PHREEQE (Ref 12) (a precursor to PHREEQC) a year later, and most GMS developed since then uses the algebraic method for determination of equilibrium states. A more comprehensive history of GMS can be found in Ref 13. Today, a large number of sophisticated GMS are available (Table 1) based on the pioneering work conducted in the 1960s and 1970s. These programs owe their existence to the compilation of thermodynamic data on which all geochemical modeling depends (Ref 14). An excellent compilation of thermodynamic and kinetic databases may be found at http://www.nea.fr/html/ science/docs/1996/nsc-doc96-27.html. Table 1 Sources of geochemical modeling software Software title AquaChem

Developer or source Waterloo Hydrogeologic, Inc. 460 Phillip Street Suite 101 Waterloo, Ontario, Canada N2L 5J2

ChemSage

www.flowpath.com/ GTT-Technologies Kaiserstraβe 100 52134 Herzogenrath, Germany

CHESS

gttserv.lth.rwth-aachen.de/gtt/ Jan van der Lee School of Mines of Paris 35, rue Saint Honoré 77305 Fontainebleau Cedex, France

EQ3/EQ6

chess.ensmp.fr Thomas J. Wolery, L-631 Lawrence Livermore National Laboratory PO Box 808

EQL/EVP, KINDIS, KIRMAT

Livermore, CA 94550, Alain Clement CNRS, Centre de Geochémie de la Surface 1 rue Blessig-67084 Strasbourg Cedex

EQS4WIN

France Mathtrek Systems 3-304 Stone Road West, Suite 165 Guelph, Ontario, Canada N1G 4W4

FactSage

www.mathtrek.com/ CRCT Ecole Polytechnique Box 6079, Station Downtown Montreal, Quebec, Canada H3C 3A7

GEMS

factsage.com/ GEMS-PSI C/O D. Kulik Waste Management Laboratory Paul Scherrer Institute CH-5232 Villigen PSI, Switzerland

GEOCHEM-PC

les.web.psi.ch/ Dr. David R. Parker

Dept. Soil and Environmental Sciences University of California Riverside, CA 92521

The Geochemist's Workbench

envisci.ucr.edu/faculty/Parker Craig Bethke Rockware, Inc. 2221 East St. #1 Golden, CO 80401

HARPHRQ

www.rockware.com A. Haworth, C.J. Tweed, and S.M. Sharland AEA Decommissioning and Radwaste Harwell Laboratory Oxon OX11 ORA, United Kingdom

HSC Chemistry

www.wiz.uni-kassel.de Outokumpu Research Oy Information Service P.O Box 60 FIN-28101 PORI, Finland

MINEQL+

www.outokumpu.com/hsc Dr. William Schecher Environmental Research Software 16 Middle Street Hallowell, ME 04347

MINTEQA2

www.mineql.com/ Center for Exposure Assessment Modeling U.S. Environmental Protection Agency Environmental Research Laboratory 960 College Station Road Athens, GA 30605-2720 www.epa.gov/ceampubl/mmedia/minteq

Visual MINTEQ PHREEQC

www.lwr.kth.se/english/OurSoftware/Vminteq/ David Parkhurst U.S. Geological Survey Box 25046, Mail Stop 418 Denver Federal Center Lakewood, CO 80225

PHRQPITZ

wwwbrr.cr.usgs.gov/projects/GWC_coupled/phreeqc U.S. Geological Survey Hydrologic Analysis Software Support Program 437 National Center Reston, VA 20192

SOILCHEM

water.usgs.gov/phrqpitz Veronica B. Lanier Sr. Lic. Officer, Physical Sciences Office of Technology Licensing University of California Berkeley, CA 94720-1620

SOLMINEQ

www.cnr.berkeley.edu/~ayang/soilchem Geochemical Applications & Modelling Software Ltd. Box 51028 Edmonton, Alberta, Canada T5W 5G5

SOLVEQ/CHILLER

www.telusplanet.net/public/geogams/index Mark H. Reed Department of Geological Sciences 1272 University of Oregon Eugene, OR 97403-1272

THERMO

[email protected] Pyrometallurgy Division, Mintek 200 Hans Strijdom Drive

Randberg, 2125, South Africa

Thermo-Calc

www.mintek.ac.za/Pyromet/Thermo Thermo_Calc Software Stockholm Technology Park Björnnäsvägen 21 SE-113 47 Stockholm, Sweden

WATEQ4F

www.thermocalc.se U.S. Geological Survey Hydrologic Analysis Software Support Program 437 National Center Reston, VA 20192

water.usgs.gov/software/wateq4f Many of these sources require a licensing or distribution fee.

References cited in this section 1. S.R. Brinkley, Jr., Calculation of the Equilibrium Composition of Systems of Many Components, J. Chem. Phys., Vol 15, 1947 p 107–110 2. R.M. Garrels, and M.E. Thompson, A Chemical Model for Sea Water at 25° and One Atmosphere Total Pressure, Am. J. Sci., Vol 260, 1962, p 57–66 3. Y.K. Kharaka and I. Barnes, “SOLMNEQ: Solution-Mineral Equilibrium Computations,” Report PB215-899, U.S. Geological Survey Computer Contributions, 1973 4. A.H. Truesdell and B.F. Jones, WATEQ, A Computer Program for Calculating Chemical Equilibria of Natural Waters, US Geol. Surv. J. Res., Vol 2, 1974, p 233–248 5. J.C. Westall, J.L. Zachary, and F.M.M. Morel, “MINEQL, A Computer Program for the Calculation of Chemical Equilibrium Composition of Aqueous Systems,” Tech. Note 18, Department of Civil Engineering, Massachusetts Institute of Technology, 1976 6. R.M. Garrels, and F.T. Mackenzie, Origin of the Chemical Compositions of Some Springs and Lakes, Equilibrium Concepts in Natural Waters, Advances in Chemistry Series 67, American Chemical Society, 1967, p 222–242 7. H.C. Hegelson, Evaluation of Irreversible Reactions in Geochemical Processes Involving Minerals and Aqueous Solutions, I: Thermodynamic Relations, Geochim. Cosmochim. Acta, Vol 32, 1968, p 853–877 8. H.C. Hegelson, A Chemical and Thermodynamic Model of Ore Deposition in Hydrothermal Systems, Miner. Soc. Am. Special Paper, Vol 3, 1970, p 155–186 9. I.K. Karpov and L.A. Kaz'min, Calculation of Geochemical Equilibria in Heterogeneous Multicomponent Systems, Geochem. Int., Vol 9, 1972, p 252–262

10. I.K. Karpov, L.A. Kaz'min, and S.A. Kashik, Optimal Programming for Computer Calculation of Irreversible Evolution in Geochemical Systems, Geochem. Int., Vol 10, 1973, p 464–470 11. T.J. Wolery, “Calculation of Chemical Equilibrium between Aqueous Solution and Minerals: the EQ3/6 Software Package,” Lawrence Livermore National Laboratory Report UCRL-52658, 1979 12. D.L. Parkhurst, D.C. Thorstenson, and L.N. Plummer, “PHREEQE—A Computer Program for Geochemical Calculations,” U.S. Geological Survey Water-Resources Investigations Report 80-96, 1980 13. C.M. Bethke, Geochemical Reaction Modeling. Concepts and Applications, Oxford University Press, 1996, 397 p 14. D.K. Nordstrom, and J.L. Munoz, Geochemical Thermodynamics, 2nd ed., Blackwell, 1994


Principles of Geochemical Modeling A model is a predictive tool for describing reality in a simplified way. For a model to be successful, it must adequately describe a real system but be simple enough that it can be easily evaluated. The system is that portion of the environment one wishes to study. A closed system has fixed composition; an open system can involve mass transfer into or out of the system. A system has boundaries (extent) that may be as small as a raindrop or as large as an ocean. A general discussion of models can be found in the article “Modeling Corrosion Processes” in this Volume. Geochemical models are normally based on the equilibrium system containing a liquid and one or more mineral phases with an initial known composition and temperature. This allows the equilibrium state of a system to be determined. In a closed system at known temperature, the model can predict the distribution of mass between mineral phases and aqueous species as well as the species activities, the fugacities of gases that may exist in the system, and the saturation state of the minerals. The constituents of the system remain in chemical equilibrium throughout the calculation. Transfer of mass into or out of the system and variation in temperature drive the system to a series of new equilibria over the course of the reaction path. The composition of a system may be buffered by equilibrium with an external gas reservoir such as the atmosphere (Fig. 3).

Fig. 3 Schematic diagram of a reaction model. Source: Ref 13 In complicated models, system composition and/or temperature may vary, and so the model must be able to account for transport of mass or heat into or out of the system. The initial equilibrium state is then the starting point for the reaction path to follow. Because geochemical models deal with equilibrium systems, it is useful to define equilibrium. In general a system is in equilibrium when no spontaneous change occurs within the defined boundaries of the system. In this case the system can be said to be in complete equilibrium, and all possible chemical reactions are in equilibrium. If one or more possible reactions tend toward equilibrium at an extremely slow rate, the system is in metastable equilibrium. In partial equilibrium, the liquid phase (fluid) may be in equilibrium but not in equilibrium with one or more mineral phases in contact with it. If a small portion of a system is chosen, this subsystem may be in local equilibrium. A successful geochemical model must at least define: • • •

Initial conditions of the system (composition and temperature) Type of equilibrium to be maintained Mass or temperature change occurring in the system during the reaction process to be modeled

The initial system contains by convention 1 kg (2.2 lb) of water, an amount that may be altered by the modeler. Mineral phases may also be included with each mineral in equilibrium with the fluid. In general, the system will be constrained by defining the initial conditions: • • • •

Solvent water mass (1 kg by default) Amounts of mineral phases Fugacities of any gases at known partial pressure Fluid chemistry, amounts of dissolved components as determined by chemical analysis

•

Activities of a species such as H+,

Once the initial conditions of a system have been so defined, the modeler can introduce a reactant (titration model) to the equilibrium system. The reactant dissolves irreversibly, which may cause mineral phases in the system to dissolve (i.e., the reactant is an acid) or the precipitation of one or more mineral phases that become saturated in relation to the fluid phase. The reaction proceeds until the reactant is exhausted or saturated in the system. In most geochemical systems, the solubility of minerals is very small, so the fluid is likely to reach saturation after only a small amount of mineral has dissolved. Many corrosion processes occur in liquids in contact with a gaseous phase. The gas (usually the atmosphere in geochemical models) acts to buffer the system chemistry by allowing the fluid to maintain equilibrium with the gas phase. Models of this type are called fixed-fugacity reaction paths. Models where the fugacities may vary during the reaction, usually because of a reaction causing the evolution of a gas, are called sliding-fugacity reaction paths.

Systems where the activities of the aqueous species may vary during the course of a reaction are called slidingactivity models and fixed-activity models when the activities remain the same. In flow-through reaction path models, the change in a fluid can be followed as it flows over or through various mineral phases, which may react with the fluid. This type of model may also be used to study changes in equilibrium in a fluid undergoing evaporation.

Reference cited in this section 13. C.M. Bethke, Geochemical Reaction Modeling. Concepts and Applications, Oxford University Press, 1996, 397 p


Limitations in Geochemical Modeling A geochemical model is only as good as the completeness and accuracy with which the initial system conditions are defined. Usually this requires a chemical analysis, the pH, and/or the oxidation state (the dissolved oxygen content) of liquid samples. The accuracy of pH in particular can be problematic in geochemical systems, as the usually low total dissolved ion content in natural systems makes pH determination difficult. In such cases the use of field pH measurements or isolation of the system from atmospheric CO 2 in transport to the lab is paramount. Accuracy may also be improved by the use of low ionic strength pH standards. Chemical analysis will usually require both atomic spectroscopy for metal cations and ion chromatography or chemometric methods for anions. Liquid samples should not be stabilized by acidification unless the fluid is filtered to remove sediment and/or colloidal solids before acidification. A model can be successful only if the thermodynamic dataset contains the mineral phases and aqueous species required for the system. Many datasets have limited ranges of pH, and so a dataset that does not contain solubility data for a species above pH 10 will be inadequate for systems at pH greater than 10. In addition, all thermodynamic datasets have errors related to chemical analysis and extrapolation. Many equilibrium constants have been determined for narrow temperature ranges, usually around room temperature, and datasets may contain extrapolated data for temperatures beyond this range. Accuracy of calculated activity coefficients is another limitation to the model. For solutions of 1 molal or less ionic strength, models that use the Debye-Huckel method are fairly accurate. At higher ionic strengths, however, a model may require “Pitzer equations” to calculate activity coefficients in brines. The program PHREEQPITZ is a GMS designed specifically for systems of this type. Most importantly, the modeler must make correct assumptions about the kind of equilibrium studied and develop an accurate perception of the reaction process. Does the system defined by the modeler accurately depict the real system? Are the mineral phases the correct ones? Are all the phases present necessary to describe the system? Is the system open or closed?


Describing Equilibrium A geochemical system for a given bulk composition can be described as the sum total of one or more phases present in the system. Phases are separable, distinct, homogeneous regions of space within the system. They are separated from each other by surfaces, which are characterized by gross changes in composition and properties between the separated phases. In most geochemical models, a system will always contain an aqueous phase containing water and dissolved species, and it may also contact mineral and gas phases. If the system has only a fluid phase, it is called homogeneous. It is called heterogeneous if it contains solid and/or gaseous phases as well. Species are the subcomponents of the phases. A gas phase may contain the species CO2 or O2, while the fluid phase may contain the species H+, Ca++, , , or Cl-. Equilibrium is calculated by determining that state (at fixed temperature and pressure) at which the phases, surfaces, and species within the system have reached the minimum value for the Gibb's free energy function. The minimum is determined by evaluating the sum of the chemical potential functions (thermodynamic components) for each of the species in the system being modeled. The problem then becomes how to choose the thermodynamic components that best describe the system. In a system to be modeled containing (a) water, (b) aqueous species, (c) minerals, and (d) gases, it is first necessary to determine all the possible independent chemical reactions possible between the species. Second, the sum total of the species present must balance the known mass of the system (i.e., if the system starts with 3 millimoles of sodium, it must have 3 millimoles of sodium at the calculated equilibrium). Third, the principle of electroneutrality requires that the ionic species in the fluid remain charge balanced. Once the distribution of the species in the fluid phase has been determined, satisfying the previously mentioned criteria, the degree of undersaturation or supersaturation of each aqueous species with respect to the mineral phases of the system can be calculated. The saturation index is the calculated saturation of a fluid with respect to a mineral phase. Undersaturated minerals have negative saturation indexes. A positive saturation index indicates the mineral is supersaturated and the system is in metastable equilibrium while the system precipitates the supersaturated mineral. The fugacity of the gases present in the system can also be calculated analogous to the saturation indices. The fugacities calculated are for a gas phase in equilibrium with the system. Finally, pe and EH, the electrochemical oxidation state of the equilibrium, may be determined. Many natural systems, especially at low temperatures, are not in equilibrium. The calculated values for pe and EH can help ascertain the disequilibrium of a system. All of the previously described calculations in the geochemical model are subject to Gibb's phase rule, better known as the degrees of freedom of the system. In most geochemical systems the degrees of freedom, or the number of pieces of information needed to describe the state, equals the sum of the number of phases in the system plus the number of aqueous species in the fluid. The number of phases equals each mineral phase plus each gas plus one for the fluid phase in which the aqueous species reside. This supplies the number of independent variables required by thermodynamics to solve each equation of the geochemical model. Geochemical modeling must also account for changes in the system during the reaction that is modeled. A good example would be the disappearance of a mineral phase that has completely dissolved in the course of the reaction. If such a change to the system takes place, the degrees of freedom for the system change. Also, a mineral surface has been eliminated, the chemical reactions in the system have changed, and the equilibrium constants have altered. The modeling software must be able to compensate for all these changes. The geochemical model describing a system consists of a set of equations with the principle unknowns typically being the mass of the water, the amount of the mineral phases, and the concentration of the aqueous species.

Although most of the equations in the model are linear, some equations are nonlinear and use the NewtonRaphson iteration method for solving the equations.


Applications of Geochemical Modeling to Corrosion As is true with all geochemical reactions in nature, corrosion is a spontaneous phenomenon. For example, iron is a common component in the crust of the earth. Being a fairly reactive metal, iron is usually found not in the elemental state but as sulfides and oxides. Structures of iron and steel exposed to nature will tend to revert to the compounds (sulfides and oxides) from which the iron was extracted. Like iron, most other metals will corrode to a greater or lesser extent when exposed to air and water in the environment. Geochemical modeling software such as Outokumpu's HSC Chemistry can provide important information on the conditions required for corrosion to occur and ways to limit or prevent corrosion. One of the earliest applications of GMS to the study of corrosion was in the computer-calculated construction of potential-pH plots, frequently referred to as Pourbaix diagrams (Ref 15). Pourbaix diagrams are a diagnostic tool to thermodynamically determine if a metal will corrode, form a passive film, or remain immune to corrosion when exposed to an aqueous environment (Fig. 4). The Pourbaix diagram consists of a plot of electrochemical potential (EH) of metallic species in water (for example Fe3O4, Fe++, Fe, etc.), versus pH (i.e. the acidity or basicity of water). The plots are constructed by considering which phase is thermodynamically most stable at a given pH, potential, and temperature.

Fig. 4 A Pourbaix diagram (potential versus pH at 25 °C) for the copper, sulfur, water system derived using HSC Chemistry Using software called POT-pH-TEMP, the Electric Power Research Institute (EPRI) (Ref 16) calculated and published a large number of Pourbaix diagrams for a series of two and three component systems from 25 to 300 °C (75 to 570 °F). Elements studied were, S, Fe, Ni, Cr, C, B, N, Cu, Ti, and Zr. The diagrams were originally applied to corrosion problems in the nuclear power industry but were of such general nature that they have been used on a wide range of corrosion problems. Much GMS will generate the data for producing Pourbaix diagrams or will generate the diagrams themselves (Fig. 4). One of the more useful features of geochemical modeling software is the ability to predict chemical speciation. In a study on the environmental effects on zinc atmospheric corrosion, Wallinder and Leygraf (Ref 17) were able to model the concentration of zinc species as a function of pH in the runoff water from zinc corrosion films (Fig. 5).

Fig. 5 Chemical speciation of zinc in precipitation runoff as a function of pH predicted by the computer models MINTEQA2 and WHAM. The data are representative of the situation immediately after the runoff is released from zinc sample surfaces exposed in the Stockholm area. Geochemical modeling software is frequently used in nuclear waste management. A number of articles have been written concerning corrosion of steel canisters and concrete containers or barriers holding nuclear waste. Atkinson et al. (Ref 18) used PHREEQE with additional data on nuclides to model corrosion of iron components and soluble species in concrete. A study of corrosion of metal canisters in a cementitious environment (high pH) using CHEQMATE and PHREEQE was done (Ref 19). The programs EQ3NR and PHREEQE were used to model the solubilities of actinide corrosion products in groundwaters (Ref 20, 21, 22). Interactions between groundwater and cement were studied to assess long-term performance of nuclear waste containment designs. The effect of the corrosion of concrete barriers and the subsequent alkaline plume that resulted were also studied (Ref 23), and others investigated corrosion of cemented waste in salt brines (Ref 24). Sharland et al. (Ref 25) used CHEQMATE and PHREEQE to study changes in EH of the pore water and the removal of oxygen by the corrosion of metal canisters at a model nuclear repository. The PHREEQE software was also used to calculate variations in EH in geochemical repositories (Ref 26). The computer model BLT-EC was developed for simulating the release and geochemical transport of contaminants from a subsurface nuclear disposal facility (Ref 27). The leaching behavior of uranium oxide subjected to groundwater corrosion and the resultant speciation was studied (Ref 28). In a later article, a thermodynamic model of the corrosion of uranium dioxide in granitic groundwater was described (Ref 29). The speciation of uranium phases in the weathering zone of the Bangombé U-Deposit using The Geochemist's Workbench (Fig. 6) was also studied (Ref 30).

Fig. 6 Programming a U4+ stability diagram (log O2 (aq) versus pH) using The Geochemist's Workbench. A number of articles have concerned geochemical modeling of the corrosion of vitrified nuclear waste or glass. The programs PHREEQE and GLASSOL were used to examine corrosion of powdered simulated high-level waste glass (Ref 31). The program CONDIMENT was used to model the fluxes of radionuclides released from the corrosion of vitrified waste into the geosphere (Ref 32). Glass corrosion in brines has been investigated using Pitzer equations (Ref 33), and the corrosion of French nuclear glass reference material R717 has also been studied (Ref 34, 35). Lolivier et al. investigated the interaction of the corrosion of R717 with boom clay (Ref 36). Diffusion of radionuclides in compacted bentonite caused by glass corrosion was modeled using PHREEQE (Ref 37). The same program was also used to model the corrosion behavior of powdered glass with or without the presence of oxygen (Ref 38). Aertens and Ghaleb looked at new methods for modeling glass corrosion and investigated two classes of methods for doing this type of modeling. (Ref 39).

In another study, geochemical modeling was used to investigate the effect of sulfate and carbonate mineralbearing reservoirs on scaling and corrosion damage of production tubing (Ref 40). The software H2OTREAT and MINTEQ were used to evaluate water treatment requirements for aquifer thermal energy storage (Ref 41). The saturation index of calcium carbonate calculated by WATEQ was used to predict corrosion and scaling in water pipe (Ref 42). Geochemical modeling was used for contamination mapping including that caused by secondary pollution from corrosion in another report (Ref 43). The PHREEQC package has been used to calculate the contributions of acid rain, dry deposition of acid gases, and dissolved CO2 (Ref 44, 45) to the dissolution of corrosion films from zinc and copper surfaces (Fig. 7). In the same studies and in one by Matthes et al. (Ref 46), PHREEQC was used to calculate solubility curves for a number of known minerals in zinc, copper, and lead corrosion films. The effect of water quality on copper corrosion using MINEQL+ was also studied (Ref 47).

Fig. 7 The contributions of strong acid (acid rain), weak acid (dissolved atmospheric CO2), and dry deposition (conversion of corrosion film minerals to more soluble species by reaction with acidic gases, SO2 and HNO3) to runoff from a copper corrosion film calculated using PHREEQC. The formation of the corrosion film mineral brochantite is also modeled.

References cited in this section 15. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, Pergamon Press, 1966 16. “Computer-Calculated Potential pH Diagrams to 300 °C,” prepared by Babcock and Wilcox Company, Electric Power Research Institute NP-3137, Vol 1–3, 1983 17. O.I. Wallinder, and C. Leygraf, Environmental Effects of Metals Induced by Atmospheric Corrosion, Outdoor Atmospheric Corrosion, STP 1421, Herb Townsend, Ed., ASTM International, p 185–199 18. A. Atkinson, F.T. Ewart, S.Y.R. Pugh, J.H. Rees, and S.M. Sharland, “Experimental and Modelling Studies of the Near-Field Chemistry for Nirex Repository Concepts, Nuclear Industry Radioactive Waste Executive,” Harwell, NSS/R-104CONF- 8711200, 1988, 14 p 19. A. Haworth, S.M. Sharland, C.J. Tweed, W. Lutz, and R.C. Ewing, Modelling of the Degradation of Cement in a Nuclear Waste Repository, Scientific Basis for Nuclear Waste Management XII, Materials Research Society, p 447–454 20. M. Snellman, “Solubility of Actinides, Fission and Corrosion Products in Simulated Groundwaters and Sievi Groundwater— PHREEQE and EQ3NR Calculations,” Valtion Teknillinen Tutkimukeskus, Otaniemi (FINLAND), YJT-90-05, 1990, 32 p

21. L.J. Criscenti and R.J. Serne, Thermodynamic Modeling of Cement/Groundwater Interaction as a Tool for Long-Term Performance Assessment, Scientific Basis for Nuclear Waste Management XIII Symposium, Materials Research Society, p 81–89 22. T. Ohe, J. Ahn, T. Ikeda, T. Kanno, T. Chiba, M. Tsukamoto, S. Nakayama, and S. Nagasaki, Analysis on Evolving Environments of Engineering Barriers of High-Level Radioactive Waste Repositories during the First 1000 Years, J. At. Energy Soc. Jpn., Vol 35 (No. 5), 1993, p 420–437 23. L. Trotignon, H. Peycelon, M. Cranga, and F. Adenot, Modelling of the Interaction between an Engineered Clay Barrier and Concrete Structures in a Deep Storage Vault, Scientific Basis for Nuclear Waste Management XXII Symposium, Vol 556, Materials Research Society, 1999, p 607–614 24. B. Kienzler, and P. Vejmelka, “Geochemical Modeling of Corrosion of Cemented Waste Forms in Salt Brines: Experimental Basis, Methods and Results,”Forschungszentrum Karlsruhe GmbH Technik and Umwelt (DE). Inst. Fuer Nukleare Entsorgungstechnik, FZKA-6046, 1998, 40 p 25. S.M. Sharland, P.W. Tasker, and C.J. Tweed, “The Evolution of Eh in the Pore Water of a Model Nuclear Waste Repository,” UKAEA Harwell Lab. (UK), Theoretical Physic Division, AERE-R— 12442, 1986, 38 p 26. K. Sasakawa, N. Gennai, J. Nakayama, R. Wada, F. Matsuda, and T. Masuda, Examination of Analysis and Evaluation Method of the Chemical Environment of Geological Repositories by Geochemical Code, Research and Development, Kobe Steel Engineering Reports, Vol 41 (No. 2), 1991, p 101–104 27. R.J. MacMinnon, Ecodynamic Reseach Associates, T.M. Sullivan, and R.R. Kinsey, BLT-EC (Breach, Leach and Transport- Equilibrium Chemistry) Data Input Guide: A Computer Model for Simulating Release and Coupled Geochemical Transport of Contaminants from a Subsurface Disposal Facility, Nuclear Regulatory Commission, 1997, 363 p 28. P. Trocellier, and J.P. Gallien, Investigation of UO2 Leaching Behavior in Groundwater Using a Nuclear Microprobe: Preliminary Results, Nucl. Instrum. Methods Phys. Res. B, Vol. 93 (No. 3), 1994, p 311–315 29. P. Trocellier, C. Cachoir, and S. Guilbert, A Simple Thermodynamical Model to Describe the Control of the Dissolution of Uranium Dioxide in Granitic Groundwater by Secondary Phase Formation, J. Nucl. Mater., Vol 256 (No. 2–3), Elsevier, 1998 p 197–206 30. K.A. Jensen, C.S. Palenik, and R.C. Ewing, U6+ Phases in the Weathering Zone of the Bangombe UDeposit: Observed and Predicted Mineralogy, Radiochim. Acta, Vol 90, 2002, p 1–9 31. J. Patyn, P. Van Iseghem, and W. Timmermans, The Long-Term Corrosion and Modeling of Two Simulated Belgian Reference High-Level Waste Glasses.II, Scientific Basis for Nuclear Waste Management XIII Symposium, Materials Research Society, 1990, p 299–307 32. E. Mouche, P. Lovera, and M. Jorda, Near- Field Modeling for the Safety Assessment of French HighLevel Waste Repositories, Proc. First-International Top Meet High Level Radioact. Waste Manage. Part 1, ASCE, Boston Society of Civil Engineers Sect, 1990, p 691–698 33. B. Grambow, and R. Muller, Chemistry of Glass Corrosion in High Saline Brines, Scientific Basis for Nuclear Waste Management XIII Symposium, Materials Research Society, 1990, p 229–240 34. L. Michaux, E. Mouche, J.C. Petit and B. Fritz, Geochemical Modeling of the Long- Term Dissolution Behavior of the French Nuclear Glass R717, Appl. Geochem., 1992, p 41–54

35. E.Y. Vernaz, and J.L. Dussossoy, Current State of Knowledge of Nuclear Waste Glass Corrosion Mechanism, The Case of R717 Glass, Appl. Geochem., 1992, p 13–22 36. P. Lolivier, K. Lemmens, and P. Iseghem, Geochemical Modeling of the Interaction of HLW Glass with Boom Clay Media, Scientific Basis for Nuclear Waste Management XXI Symposium I, Materials Research Society, 1998, p 399–406 37. M. Tsukamoto, T. Ohe, T. Fujita, R. Hesbol, and H.P. Hermansson, Diffusion of Radionuclides in Compacted Bentonite: Results from Combined Glass Dissolution and Migration Tests, Scientific Basis for Nuclear Waste Management XVII Symposium I, Materials Research Society, 1995, p 291–298 38. Y. Inagaki, H. Furuya, K. Idemitsu, T. Marda, A. Sakai, T. Banba, and S. Muraoka, Corrosion Behavior of a Powdered Simulated Nuclear Waste Glass under Anoxic Condition, Scientific Basis for Nuclear Waste Management XVII Symposium I, Materials Research Society, 1995, p 23–30 39. M. Aertens, and D. Ghaleb, New Techniques for Modeling Glass Dissolution, J. Nuclear Materials, Vol 298 (No.1–2), 2001, p 37–46 40. D.B. Macgowan, T.L. Dunn, and R.C. Surdam, Geochemical Modeling of Scale Formation and Formation Damage during Production from Sulfate and Carbonate Mineral-Bearing Reservoirs, Am. Assoc. Petrol. Geolog., Vol 75 (No. 3), 1991, p 626 41. L.W. Vail, E.A. Jence, and L.E. Eary, H2OTREAT: An Aid for Evaluating Water Treatment Requirements for Aquifer Thermal Energy Storage, Vol 4, Society of Automotive Engineering, 1992, p 4.131–4.135 42. L.E. Marin, E.P. Santa Anna, and G.V. Oliman, Application of Geochemical Modeling in Hydraulic Engineering, Ing. Hidraul. Mex., Vol 9, 1994, p 63–69 43. T.S. Lee, S.K. Lee, and Y.K. Hong, Environmental Geophysics and Geochemistry for Contamination Mapping and Monitoring I, Korea Institute of Geology and Mining and Materials, KR-95-T-3, 1995, 439 p 44. S.D. Cramer, S.A. Matthes, G.R. Holcomb, B.S. Covino Jr., and S.J. Bullard, “Precipitation Runoff and Atmospheric Corrosion,” Paper 00452, presented at Corrosion/2000, (Houston TX), NACE International, 2000, 15 p 45. S.D. Cramer, S.A. Matthes, B.S. Covino Jr., S.J. Bullard, and G.R. Holcomb, Environmental Factors Affecting the Atmospheric Corrosion of Copper, Outdoor Atmospheric Corrosion, STP 1421, H.E. Townsend, Ed., ASTM International, 2002, p 245–264 46. S.A. Matthes, S.D. Cramer, B.S. Covino Jr., S.J. Bullard, and G.R. Holcomb, Precipitation Runoff from Lead, Outdoor Atmospheric Corrosion, STP 1421, H.E. Townsend, Ed., ASTM International, 2002, p 265–274 47. S.O. Pehkoen, A. Palit, and X. Zhang, Effect of Specific Water Quality Parameters on Copper Corrosion, Corrosion, Vol 58 (No. 2), 2002, p 156–165


Geochemical Modeling Features Geochemical modeling can perform many useful tasks for the corrosion scientist: • • • • • • • • • • • •

Construction of EH vs pH (Pourbaix) diagrams Speciation in aqueous and nonaqueous solutions (Fig. 8) Construction of solubility curves Equilibrium in high-temperature corrosion systems Metals complexing Calculation of equilibrium concentrations Reaction path calculations Gas-liquid-solid phase calculations (Fig. 9) Surface complexation Inverse modeling Automatic reaction balancing Modeling irreversible reactions

Fig. 8 SOLMINEQ, a geochemical modeling program. Pictured here is the module for aqueous speciation.

Fig. 9 High-temperature phase diagram for the Fe-Cr-Ni-C system using FactSage (GTT Technologies) geochemical modeling software These calculations are performed by the use of thermodynamics equations and large databases of thermodynamic data. A number of geochemical modeling software packages that can do these calculations are available, both as commercial and freeware products. Some of these programs are quite complex and have a steep learning curve. Some may require additional user training.


References 1. S.R. Brinkley, Jr., Calculation of the Equilibrium Composition of Systems of Many Components, J. Chem. Phys., Vol 15, 1947 p 107–110 2. R.M. Garrels, and M.E. Thompson, A Chemical Model for Sea Water at 25° and One Atmosphere Total Pressure, Am. J. Sci., Vol 260, 1962, p 57–66

3. Y.K. Kharaka and I. Barnes, “SOLMNEQ: Solution-Mineral Equilibrium Computations,” Report PB215-899, U.S. Geological Survey Computer Contributions, 1973 4. A.H. Truesdell and B.F. Jones, WATEQ, A Computer Program for Calculating Chemical Equilibria of Natural Waters, US Geol. Surv. J. Res., Vol 2, 1974, p 233–248 5. J.C. Westall, J.L. Zachary, and F.M.M. Morel, “MINEQL, A Computer Program for the Calculation of Chemical Equilibrium Composition of Aqueous Systems,” Tech. Note 18, Department of Civil Engineering, Massachusetts Institute of Technology, 1976 6. R.M. Garrels, and F.T. Mackenzie, Origin of the Chemical Compositions of Some Springs and Lakes, Equilibrium Concepts in Natural Waters, Advances in Chemistry Series 67, American Chemical Society, 1967, p 222–242 7. H.C. Hegelson, Evaluation of Irreversible Reactions in Geochemical Processes Involving Minerals and Aqueous Solutions, I: Thermodynamic Relations, Geochim. Cosmochim. Acta, Vol 32, 1968, p 853–877 8. H.C. Hegelson, A Chemical and Thermodynamic Model of Ore Deposition in Hydrothermal Systems, Miner. Soc. Am. Special Paper, Vol 3, 1970, p 155–186 9. I.K. Karpov and L.A. Kaz'min, Calculation of Geochemical Equilibria in Heterogeneous Multicomponent Systems, Geochem. Int., Vol 9, 1972, p 252–262 10. I.K. Karpov, L.A. Kaz'min, and S.A. Kashik, Optimal Programming for Computer Calculation of Irreversible Evolution in Geochemical Systems, Geochem. Int., Vol 10, 1973, p 464–470 11. T.J. Wolery, “Calculation of Chemical Equilibrium between Aqueous Solution and Minerals: the EQ3/6 Software Package,” Lawrence Livermore National Laboratory Report UCRL-52658, 1979 12. D.L. Parkhurst, D.C. Thorstenson, and L.N. Plummer, “PHREEQE—A Computer Program for Geochemical Calculations,” U.S. Geological Survey Water-Resources Investigations Report 80-96, 1980 13. C.M. Bethke, Geochemical Reaction Modeling. Concepts and Applications, Oxford University Press, 1996, 397 p 14. D.K. Nordstrom, and J.L. Munoz, Geochemical Thermodynamics, 2nd ed., Blackwell, 1994 15. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, Pergamon Press, 1966 16. “Computer-Calculated Potential pH Diagrams to 300 °C,” prepared by Babcock and Wilcox Company, Electric Power Research Institute NP-3137, Vol 1–3, 1983 17. O.I. Wallinder, and C. Leygraf, Environmental Effects of Metals Induced by Atmospheric Corrosion, Outdoor Atmospheric Corrosion, STP 1421, Herb Townsend, Ed., ASTM International, p 185–199 18. A. Atkinson, F.T. Ewart, S.Y.R. Pugh, J.H. Rees, and S.M. Sharland, “Experimental and Modelling Studies of the Near-Field Chemistry for Nirex Repository Concepts, Nuclear Industry Radioactive Waste Executive,” Harwell, NSS/R-104CONF- 8711200, 1988, 14 p 19. A. Haworth, S.M. Sharland, C.J. Tweed, W. Lutz, and R.C. Ewing, Modelling of the Degradation of Cement in a Nuclear Waste Repository, Scientific Basis for Nuclear Waste Management XII, Materials Research Society, p 447–454

20. M. Snellman, “Solubility of Actinides, Fission and Corrosion Products in Simulated Groundwaters and Sievi Groundwater— PHREEQE and EQ3NR Calculations,” Valtion Teknillinen Tutkimukeskus, Otaniemi (FINLAND), YJT-90-05, 1990, 32 p 21. L.J. Criscenti and R.J. Serne, Thermodynamic Modeling of Cement/Groundwater Interaction as a Tool for Long-Term Performance Assessment, Scientific Basis for Nuclear Waste Management XIII Symposium, Materials Research Society, p 81–89 22. T. Ohe, J. Ahn, T. Ikeda, T. Kanno, T. Chiba, M. Tsukamoto, S. Nakayama, and S. Nagasaki, Analysis on Evolving Environments of Engineering Barriers of High-Level Radioactive Waste Repositories during the First 1000 Years, J. At. Energy Soc. Jpn., Vol 35 (No. 5), 1993, p 420–437 23. L. Trotignon, H. Peycelon, M. Cranga, and F. Adenot, Modelling of the Interaction between an Engineered Clay Barrier and Concrete Structures in a Deep Storage Vault, Scientific Basis for Nuclear Waste Management XXII Symposium, Vol 556, Materials Research Society, 1999, p 607–614 24. B. Kienzler, and P. Vejmelka, “Geochemical Modeling of Corrosion of Cemented Waste Forms in Salt Brines: Experimental Basis, Methods and Results,”Forschungszentrum Karlsruhe GmbH Technik and Umwelt (DE). Inst. Fuer Nukleare Entsorgungstechnik, FZKA-6046, 1998, 40 p 25. S.M. Sharland, P.W. Tasker, and C.J. Tweed, “The Evolution of Eh in the Pore Water of a Model Nuclear Waste Repository,” UKAEA Harwell Lab. (UK), Theoretical Physic Division, AERE-R— 12442, 1986, 38 p 26. K. Sasakawa, N. Gennai, J. Nakayama, R. Wada, F. Matsuda, and T. Masuda, Examination of Analysis and Evaluation Method of the Chemical Environment of Geological Repositories by Geochemical Code, Research and Development, Kobe Steel Engineering Reports, Vol 41 (No. 2), 1991, p 101–104 27. R.J. MacMinnon, Ecodynamic Reseach Associates, T.M. Sullivan, and R.R. Kinsey, BLT-EC (Breach, Leach and Transport- Equilibrium Chemistry) Data Input Guide: A Computer Model for Simulating Release and Coupled Geochemical Transport of Contaminants from a Subsurface Disposal Facility, Nuclear Regulatory Commission, 1997, 363 p 28. P. Trocellier, and J.P. Gallien, Investigation of UO2 Leaching Behavior in Groundwater Using a Nuclear Microprobe: Preliminary Results, Nucl. Instrum. Methods Phys. Res. B, Vol. 93 (No. 3), 1994, p 311–315 29. P. Trocellier, C. Cachoir, and S. Guilbert, A Simple Thermodynamical Model to Describe the Control of the Dissolution of Uranium Dioxide in Granitic Groundwater by Secondary Phase Formation, J. Nucl. Mater., Vol 256 (No. 2–3), Elsevier, 1998 p 197–206 30. K.A. Jensen, C.S. Palenik, and R.C. Ewing, U6+ Phases in the Weathering Zone of the Bangombe UDeposit: Observed and Predicted Mineralogy, Radiochim. Acta, Vol 90, 2002, p 1–9 31. J. Patyn, P. Van Iseghem, and W. Timmermans, The Long-Term Corrosion and Modeling of Two Simulated Belgian Reference High-Level Waste Glasses.II, Scientific Basis for Nuclear Waste Management XIII Symposium, Materials Research Society, 1990, p 299–307 32. E. Mouche, P. Lovera, and M. Jorda, Near- Field Modeling for the Safety Assessment of French HighLevel Waste Repositories, Proc. First-International Top Meet High Level Radioact. Waste Manage. Part 1, ASCE, Boston Society of Civil Engineers Sect, 1990, p 691–698 33. B. Grambow, and R. Muller, Chemistry of Glass Corrosion in High Saline Brines, Scientific Basis for Nuclear Waste Management XIII Symposium, Materials Research Society, 1990, p 229–240

34. L. Michaux, E. Mouche, J.C. Petit and B. Fritz, Geochemical Modeling of the Long- Term Dissolution Behavior of the French Nuclear Glass R717, Appl. Geochem., 1992, p 41–54 35. E.Y. Vernaz, and J.L. Dussossoy, Current State of Knowledge of Nuclear Waste Glass Corrosion Mechanism, The Case of R717 Glass, Appl. Geochem., 1992, p 13–22 36. P. Lolivier, K. Lemmens, and P. Iseghem, Geochemical Modeling of the Interaction of HLW Glass with Boom Clay Media, Scientific Basis for Nuclear Waste Management XXI Symposium I, Materials Research Society, 1998, p 399–406 37. M. Tsukamoto, T. Ohe, T. Fujita, R. Hesbol, and H.P. Hermansson, Diffusion of Radionuclides in Compacted Bentonite: Results from Combined Glass Dissolution and Migration Tests, Scientific Basis for Nuclear Waste Management XVII Symposium I, Materials Research Society, 1995, p 291–298 38. Y. Inagaki, H. Furuya, K. Idemitsu, T. Marda, A. Sakai, T. Banba, and S. Muraoka, Corrosion Behavior of a Powdered Simulated Nuclear Waste Glass under Anoxic Condition, Scientific Basis for Nuclear Waste Management XVII Symposium I, Materials Research Society, 1995, p 23–30 39. M. Aertens, and D. Ghaleb, New Techniques for Modeling Glass Dissolution, J. Nuclear Materials, Vol 298 (No.1–2), 2001, p 37–46 40. D.B. Macgowan, T.L. Dunn, and R.C. Surdam, Geochemical Modeling of Scale Formation and Formation Damage during Production from Sulfate and Carbonate Mineral-Bearing Reservoirs, Am. Assoc. Petrol. Geolog., Vol 75 (No. 3), 1991, p 626 41. L.W. Vail, E.A. Jence, and L.E. Eary, H2OTREAT: An Aid for Evaluating Water Treatment Requirements for Aquifer Thermal Energy Storage, Vol 4, Society of Automotive Engineering, 1992, p 4.131–4.135 42. L.E. Marin, E.P. Santa Anna, and G.V. Oliman, Application of Geochemical Modeling in Hydraulic Engineering, Ing. Hidraul. Mex., Vol 9, 1994, p 63–69 43. T.S. Lee, S.K. Lee, and Y.K. Hong, Environmental Geophysics and Geochemistry for Contamination Mapping and Monitoring I, Korea Institute of Geology and Mining and Materials, KR-95-T-3, 1995, 439 p 44. S.D. Cramer, S.A. Matthes, G.R. Holcomb, B.S. Covino Jr., and S.J. Bullard, “Precipitation Runoff and Atmospheric Corrosion,” Paper 00452, presented at Corrosion/2000, (Houston TX), NACE International, 2000, 15 p 45. S.D. Cramer, S.A. Matthes, B.S. Covino Jr., S.J. Bullard, and G.R. Holcomb, Environmental Factors Affecting the Atmospheric Corrosion of Copper, Outdoor Atmospheric Corrosion, STP 1421, H.E. Townsend, Ed., ASTM International, 2002, p 245–264 46. S.A. Matthes, S.D. Cramer, B.S. Covino Jr., S.J. Bullard, and G.R. Holcomb, Precipitation Runoff from Lead, Outdoor Atmospheric Corrosion, STP 1421, H.E. Townsend, Ed., ASTM International, 2002, p 265–274 47. S.O. Pehkoen, A. Palit, and X. Zhang, Effect of Specific Water Quality Parameters on Copper Corrosion, Corrosion, Vol 58 (No. 2), 2002, p 156–165


Selected References • • • • • • • • • • • •

F. Albaredo, Introduction to Geochemical Modeling, Cambridge University Press, April 1996, 563 p G.M. Anderson, Thermodynamics of Natural Systems, John Wiley & Sons, 1996, 382 p G.M. Anderson, and D.A. Crerar, Thermodynamics in Geochemistry: The Equilibrium Model, Oxford University Press, 1993, 588 p C.A.J. Appelo and D. Postma, Geochemistry, Groundwater and Pollution, Balkema, 1994, 536 p C.M. Bethke, Geochemical Reaction Modeling: Concepts and Applications, Oxford University Press, 1996, 397 p K. Denbigh, The Principles of Chemical Equilibrium, 4th ed., Cambridge University Press, 1981, 494 p J.I. Drever, The Geochemistry of Natural Waters, 2nd ed. Prentice Hall, 1988, 437 p P. Fletcher, Chemical Thermodynamics for Earth Scientists, Longman/Harlow, 1993, 464 p D.K. Nordstrom, and J.L. Munoz, Geochemical Thermodynamics, 2nd ed., Blackwell Scientific Publications, 1994, 493 p J.F. Pankow, Aquatic Chemistry Concepts, Lewis Publishers, 1991, 683 p K.S. Pitzer, Thermodynamics, McGraw-Hill, 1995, 626 p W. Stumm, and J.J. Morgan, Aquatic Chemistry, 3rd ed., Wiley-Interscience, 1996, 1022 p

D.W. Shoesmith, Kinetics of Aqueous Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 42–51

Kinetics of Aqueous Corrosion D.W. Shoesmith, University of Western Ontario

Introduction THIS ARTICLE gives a general introduction to the kinetics of aqueous corrosion, with an emphasis on electrochemical principles. The concept of a corrosion process as a combination of electrochemical reactions is important, because, for aqueous corrosion processes, electrochemical techniques are the predominant methods used in their study to measure rates as well as in the development of on-line sensors. The primary goal of this article is to introduce the electrochemical concepts. Readers interested in more fundamental details of electrochemistry, electrochemical methods, and their application to corrosion processes will find Ref 1, 2, 3, 4 useful.

References cited in this section 1. J.H. West, Electrodeposition and Corrosion Processes, Van Nostrand Reinhold, 1971 2. D. Jones, Principles and Prevention of Corrosion, 2nd ed., Prentice Hall, 1996

3. L.L. Shreir, R.A. Jarman, and G.T. Burstein, Ed., Metal/Environment Interactions, Corrosion, Vol 1, 3rd ed., Butterworth/Heinemann, 1994 4. L.L. Shreir, R.A. Jarman, and G.T. Burstein, Ed., Corrosion Control, Corrosion, Vol 2, 3rd ed., Butterworth/Heinemann, 1994

D.W. Shoesmith, Kinetics of Aqueous Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 42–51 Kinetics of Aqueous Corrosion D.W. Shoesmith, University of Western Ontario

Basis of Corrosion Reactions Aqueous corrosion is an electrochemical process occurring at the interface between a material (commonly, but not exclusively, a metal) and an aqueous solution. For corrosion to occur, an oxidation reaction (generally, metal dissolution and/or metal oxide formation) and a reduction reaction (commonly, proton, water, or dissolved oxygen reduction) must occur simultaneously. In electrochemical terms, an anodic (oxidation, or Ox) reaction: M → Mn+ + ne-

(Eq 1) n+

where M denotes a metal and M a dissolved metal cation, is coupled to a cathodic (reduction, or R) reaction: Ox + ne- → Red

(Eq 2)

where Ox denotes a soluble oxidant and Red the reduced form of the oxidant that may or may not be a soluble species. An example of a cathodic reagent that produces a soluble product is oxygen reduction: O2 + 2H2O + 4e- → 4OH-

(Eq 3)

An example of a cathodic reaction producing a gaseous product is proton reduction: 2H+ + 2e- → H2

(Eq 4)

The sum of two electrochemical half-reactions, one anodic (such as Eq 1) and one cathodic (such as Eq 2), is the overall corrosion reaction: M + Ox → Mn+ + Red

(Eq 5)

An example is the dissolution of iron in an aerated solution: 2Fe + O2 + 2H2O → 2Fe2+ + 4OH-

(Eq 6)

Thus, corrosion is the coupling together of two electrochemical reactions on the same surface. This coupling occurs at a single potential, known as the corrosion potential (Ecorr). This potential will depend on the relative rates of the coupled anodic and cathodic reactions (see subsequent information), and the oxidation state of the dissolved metal cation may change as Ecorr changes. Thus, an oxidant capable of driving the corrosion potential to a more positive value could produce a higher oxidation state of the metal (e.g., Fe3+ rather than Fe2+). The corroding material-solution combination can be considered a short-circuited galvanic cell in which the energy is dissipated by the consumption of cathodic reagent (oxidant). This situation is illustrated schematically in Fig. 1(a).

Fig. 1 Galvanic cells. (a) Schematic illustrating the short-circuit galvanic cell that exists during corrosion. (b) The coupling of an anodic reaction with two distinct cathodic reactions. The relative anodic (Aa) and cathodic (Ac) areas of the corroding surface are also illustrated. To maintain a balance, the amount of cathodic reagent (Ox) consumed must be equal to the amount of corrosion product (Mn+) formed. Because electrons are liberated by the anodic reaction and consumed by the cathodic reaction, corrosion can be expressed in terms of an electrochemical current (I). Furthermore, the requirement for mass balance requires that the current flowing into the cathodic reaction must be equal to the current flowing out of the anodic reaction. Clearly, by inspection of Fig. 1(a), these currents are opposite in sign. By definition, under open-circuit or freely corroding conditions: Ia = |Ic| = Icorr

(Eq 7)

where Ia is the anodic current, Ic is the cathodic current, and Icorr is the corrosion current. The short-circuited nature of the overall corrosion process means that Icorr cannot be directly measured on open circuit. Techniques for its measurement are discussed elsewhere in this Volume. The value of Icorr is a measure of the rate of the corrosion process and therefore of the rate of material degradation. The current and the amount of material corroded are related by Faraday's law: (Eq 8) where Icorr is expressed in amps; t is the time of exposure to the corrosive environment (seconds); nF is the number of coulombs (C) required to convert 1 mol of material to corrosion product; n is the number of electrons transferred or liberated in the oxidation reaction; F is the Faraday constant (96,480 C/mol); M is the molecular weight of the material in grams (g); and w is the mass of corroded material (g). It is possible for the anodic reaction to be supported by more than one cathodic reaction (Fig. 1b). For example, in oxygenated acidic solutions, the generic corrosion reaction (Eq 1) could be driven by both proton reduction (Eq 4) and oxygen reduction (Eq 3). When complex alloys are involved, the anodic corrosion reaction may also be the sum of more than one dissolution reaction; that is, the congruent dissolution of a Ni-Cr-Mo alloy would involve the dissolution of each alloy component with partial anodic currents proportional to the atomic fraction of each component in the alloy. The corrosion current (Icorr) then equals the sum of the component partial currents: Icorr = ∑Ia = -∑Ic

(Eq 9)

Additionally, the area and location of the anodic and cathodic sites (Aa and Ac, Fig. 1b) may be different. As a consequence, although the total anodic and cathodic currents must be equal, the respective current densities need not be: Ia = -Ic; Aa ≠ Ac And, therefore:

(Eq 10)

(Eq 11) The term I/A is a current density and will be designated i. This inequality in current densities can have serious implications. For a smooth, single-component metal surface, the anodic and cathodic sites will be separated, at any one instant, by only a few nanometers. In general terms, the anodic and cathodic sites will shift with time, so that the surface reacts evenly as it undergoes general corrosion. At the atomic scale, however, the surface is not necessarily smooth and is usually considered to comprise surface terraces, ledges, and kinks, as illustrated in Fig. 2. Also, surface defects such as ad-atoms, vacancies, and emergent defects may also exist. Because the coordination number (N) of atoms in these various locations differs, that is: Nterrace > Nledge > Nkink

(Eq 12)

there will be a difference in strength of surface bonding, and atoms will be preferentially removed in the order kink, then ledge, then terrace; that is, surface defects represent potential anodic sites.

Fig. 2 A metal surface on the atomic scale showing the existence of kinks, ledges, and terraces The separation of anodes and cathodes induced by these atomic defects is, however, minor compared to the separations induced by other surface features, such as surface asperity, alloy phases, grain boundaries, impurity inclusions, residual stresses due to processes such as cold working, and high-resistance surface oxide films. These features can often lead to the stabilization of discrete anodic and cathodic sites. The specific combination of a small anode area and a large cathode area can lead to large discrepancies in anodic and cathodic current densities, that is, to large anodic current densities at a small number of discrete sites. Such a situation is particularly dangerous when the majority of a metal surface is oxide covered (acting as a cathode) and only a small number of bare metal sites (acting as anodes) are exposed to the solution environment. Such a situation exists during pitting, crevice corrosion, and stress-corrosion cracking, as illustrated for pitting in Fig. 3.

Fig. 3 A small anode/large cathode situation that can exist at a local corrosion site. Aa and Ac are the available anode and cathode areas; Mn+ is the corrosion product Thus, it is clear that aqueous corrosion is a complicated process that can occur in a wide variety of forms and is affected by many chemical, electrochemical, and metallurgical variables, including: • • • •

The composition and metallurgical properties of the metal or alloy The chemical composition and physical properties of the environment, such as temperature and conductivity The presence or absence of surface films The properties of the surface films, such as resistivity, thickness, nature of defects, and coherence


Thermodynamic Basis for Corrosion The thermodynamic feasibility of a particular corrosion reaction is determined by the relative values of the equilibrium potentials, Ee, of the electrochemical reactions involved. These potentials can be determined from the Nernst equation, which, for the two half-reactions, Eq 1 and Eq 2, can be written: (Eq 13) where the activity of the solid, M, is taken as 1, and: (Eq 14) to yield for the overall corrosion reaction: (Eq 15) where ΔE0 is the difference in standard potentials for the anodic (Eq 1) and cathodic (Eq 2) reactions (available from a table of standard potentials), R is the gas constant equal to 8.314 J/ mol · K, T is absolute temperature, n and F are as defined in Eq 8, and ai represents the activities of the various species involved. Because the activity of a solid is 1, aM is eliminated from Eq 15, and it is more common to write the equation in terms of concentration (c): (Eq 16) The thermodynamics of a particular metal- aqueous system can be summarized in a potential-pH, or Pourbaix diagram, as described elsewhere in this Volume. The key feature in determining whether or not a particular corrosion reaction can proceed is the difference in equilibrium potentials (Ee) for the two component electrochemical half-reactions, as illustrated in Fig. 4. That is: (Eq 17) Thus, the thermodynamic driving force (ΔEtherm) for corrosion is given by:

(Eq 18)

Fig. 4 The thermodynamic driving force for corrosion across a metal/aqueous solution interface in the presence of a soluble oxidant, Ox. The values of the equilibrium potentials are shown schematically. It is possible that this driving force could increase or decrease with time if either the concentration of available oxidant, cOx, or the concentration of soluble corrosion product, , change with time, as illustrated in Fig. 4 and discernible by inspection of Eq 16. However, the most important questions for the corrosion engineer are: How fast does the corrosion reaction occur, and can it be prevented or at least slowed to an acceptable rate? To answer these questions and to determine a course of action, it is essential to have some knowledge of the steps involved in the overall corrosion process.


Kinetics of Aqueous Corrosion Processes

The overall process can be controlled by any one or combination of several reactions depicted in Fig. 5. The interfacial kinetics of either the anodic (step 1) or cathodic (step 2) reactions could be rate controlling. Alternatively, if these reactions are fast and the concentration of cathodic reagent (cOx) is low, then the rate of transport of the oxidant (Ox) to the cathodic site (step 3 → 2) could be rate limiting. This situation is quite common for corrosion driven by dissolved oxygen that has limited solubility. If the metal dissolution reaction (step 1) is reversible, that is, the reverse metal deposition (Mn+ + ne- → M) can also occur, then the rate of transport of Mn+ away from the anodic site (step 4) could be rate determining.

Fig. 5 Various possible reaction steps on a corroding metal surface. Reactions 1 to 8 are defined in the text. The presence of corrosion films can add many additional complications. If the anodic reaction (step 1) is fast and the transport step (step 4) is slow, then local supersaturation with dissolved corrosion product at the corroding surface: (Eq 19) could lead to the deposition of corrosion product deposits such as oxides, hydroxides, or metal salts. These deposition processes could be accelerated or prevented by local pH changes due to the cathodic reaction (Eq 3) or metal cation hydrolysis equilibria: (Eq 20) The local situation at the corroding surface becomes complicated, and the primary parameter governing whether or not corrosion product deposits form, aside from the balance between interfacial kinetics and solution transport, is the deposit solubility and how it varies with pH (Fig. 6). As shown, the solubility tends to be at a minimum at approximately neutral pH and to increase at high and, especially, low pH. At high pH, solubility is increased by the stabilization of hydrolyzed metal cations (i.e., the equilibrium the reaction in Eq 20 is pushed to the right), whereas in acidic solutions, the oxide/hydroxide is destabilized by the increased proton concentration (i.e., the equilibrium in Eq 20 is pushed to the left). This is a very simplified description of oxide/hydroxide solubility, and the reader is referred to Ref. 5 for a more detailed discussion.

Fig. 6 Solubility of a metal cation as a function of solution pH Precipitation reactions of this type are likely to produce porous deposits, and corrosion could then become controlled by the transport of Mn+ or Ox (step 6 in Fig. 5) through these porous deposits. By contrast, when coherent, nonporous oxide films (passive films) form spontaneously on the metal surface by solid state as opposed to precipitation reactions, then ionic or defect transport processes through the oxide (step 7 in Fig. 5) would ensure extremely low corrosion rates. This represents the condition of passivity. The presence of defects in the passive film in the form of pores, grain boundaries, or fractures can lead to localized corrosion. Finally, it is possible for the corrosion process to be controlled by the electronic conductivity of passive films (step 8 in Fig. 5) when the cathodic reaction occurs on the surface of the film. Using this range of possibilities, the remainder of this article discusses some of these processes and the laws that govern them.

Reference cited in this section 5. C.F. Baes and R.E. Mesmer, The Hydrolysis of Cations, John Wiley and Sons, 1976.


Activation Control of Corrosion Activation control is the term used to describe control of the corrosion process by the electrochemical halfreactions (Eq 1, 2) and step 1 and 2 in Fig. 5 The overall anodic reaction is the transfer of a metal atom from the metal surface to the aqueous solution as the cation Mn+ or as some hydrolyzed complexed

or anion-

form, where complexation is with the anion Ax-: (Eq 21)

As indicated in Eq 21, various surface intermediates may be involved in the overall anodic dissolution process, although these intermediates may only be detectable in a detailed electrochemical investigation. Thus, the

simplest proposed mechanism for the anodic dissolution of iron under acidic conditions proceeds via the following steps: Fe + H2O ↔ Fe(OH)ads + H+ + e

(Eq 22)

Fe(OH)ads ↔ FeOH+ + e-

(Eq 23)

FeOH+ + H+ ↔ Fe2+ + H2O

(Eq 24)

Similarly, the cathodic reaction, for example, O2 reduction, proceeds via a sequence of reaction steps involving intermediates such as hydrogen peroxide: O2 + 2H+ + 2e- → H2O2

(Eq 25)

H2O2 + 2H+ + 2e- → 2H2O

(Eq 26)

Either the anodic or cathodic reaction can control the overall rate of the corrosion reaction. Electrochemical methods (Ref 6) can be used to disturb these reactions from their equilibrium potentials and hence to determine the relationship between current and potential. An example of the form of this relationship, demonstrating a metal dissolution/deposition reaction, is shown in Fig. 7. This curve follows the Butler-Volmer equation: (Eq 27) where I is the current; I0 is the exchange current; F is the Faraday constant; R is the gas constant; T the absolute temperature; and α is the transfer or symmetry coefficient, generally taken to be 0.5. The term η is the overpotential, defined by: (Eq 28) where E is the experimentally applied potential, and Ee is given by the Nernst equation (Eq 16), with ΔE0 = . The overpotential is a measure of how far the reaction is from equilibrium.

Fig. 7 Current-potential (Butler-Volmer) relationship for a metal dissolution/deposition reaction. The solid line shows the measurable current; the dashed lines (Ia, Ic) show the partial dissolution/deposition

currents; Io is the exchange current at Ee; (ηa) is the anodic overpotential that would exist if the potential were held at point 1. At equilibrium (η = 0; E = Ee), no measurable current flows, but the equilibrium is dynamic, with the rate of metal dissolution (M → Mn+ + ne-) (Ia) equal to the rate of metal cation deposition (Mn+ + ne- → M) (-Ic): Ia = -Ic = Io

(Eq 29)

where Io is termed the exchange current for this electrochemical reaction. The exchange current is the equivalent of an electrochemical rate constant and is therefore a direct measure of the kinetics of the reaction. The dynamic nature of the reaction at E = Ee can be appreciated by consideration of the partial anodic and cathodic currents, plotted as dashed lines in Fig. 7. If the potential is made more positive than the equilibrium potential, then Ia > |Ic| (i.e., the first exponential term in Eq 27 increases in value, and the second one decreases), and the metal dissolution reaction will proceed. Similarly, if E is made more negative than the equilibrium potential, Ia > |Ic|, then metal cation deposition proceeds. Over a short potential range, the two reactions oppose each other, but for sufficiently large overpotentials (ηa, anodic; or ηc, cathodic), one reaction occurs at a negligible rate; that is, one or the other of the exponential terms in the Butler-Volmer equation (Eq 27) becomes negligible. In this condition, the overpotential is in the Tafel region (as indicated by point 1 in Fig. 7), and for this positive polarization, the metal dissolution current is given by: (Eq 30) or (Eq 31) and a plot of log Ia versus ηa will yield an intercept from which Io can be obtained and will have a slope with a Tafel coefficient, βa, given by: (Eq 32) A similar procedure could be applied for a sufficiently large cathodic overpotential (ηc) to yield an equal value of Io as well as a value for the cathodic Tafel coefficient, βc. The values of βa and βc contain much information on the mechanism of the dissolution/deposition reaction steps but are commonly used as empirical coefficients in corrosion engineering studies. A similar analysis could be performed for a potential cathodic reaction (Eq 2), and Fig. 8 shows the two current-potential curves on the same plot.

Fig. 8 Current-potential relationships for a metal dissolution/deposition process (M ↔ Mn+ + e-) and an oxidant/reductant reaction (O + ne- ↔ R) showing the coupling together of the anodic component of one

reaction to the cathodic component of the other to yield a corrosion reaction proceeding at the corrosion potential, Ecorr The establishment of a corrosion reaction involves the coupling together of the anodic half of one reaction to the cathodic half of another. Which reaction provides which half is determined by the relative values of the equilibrium potentials, that is, condition (Eq 17), as illustrated in Fig. 4. This coupling of half-reactions produces an equation relating the measured current to the corrosion current, which is similar in form to the Butler-Volmer equation (Eq 27): (Eq 33) where i and icorr represent current densities (current/surface area), αa ≠ αc, and there is no reason why αa + αc = 1, as was the case with the Butler-Volmer equation. Note: From this point on, current densities (i) rather than currents (I) will be used. Because this coupling of half-reactions produces a short-circuited corrosion reaction on the surface of the metal, the anodic current due to metal dissolution must be equal and opposite in sign to the cathodic current due to oxidant reduction: ia = -ic = icorr

(Eq 34)

where ia is given by the first term in Eq 33 and ic by the second. This coupling together to produce equal anodic and cathodic currents can only occur at a single potential, designated the corrosion potential, Ecorr, which must lie between the two equilibrium potentials (Fig. 8), thus satisfying the condition in Eq 15: (Ee)a < Ecorr < (Ee)c

(Eq 35)

where (Ee)a is equivalent to and (Ee)c to (Ee)Ox/Red. The metal dissolution (anodic) reaction is driven by an anodic activation overpotential: (Eq 36) and the oxidant reduction (cathodic) reaction by a cathodic activation overpotential: (Eq 37) Obviously, from Eq 18: (Eq 38) Two additional observations can be made with regard to Fig. 8. First, ΔE is sufficiently large that Ecorr is in the Tafel regions for both the anodic and cathodic half-reaction. This is not necessarily always the case, as discussed subsequently. Second, the two current-potential curves are not necessarily symmetrical and seldom have an identical shape. The shape and symmetry of the curves are determined by the differences in io for the coupled anodic and cathodic reactions and the values of the Tafel coefficients, βa and βc. In Fig. 8, the metal dissolution/deposition reaction is shown to have a larger io than the oxidant/reductant reaction, and the anodic and cathodic branches are shown close to symmetrical (i.e., βa ~ βc). The consequence of the large io is that the are required to achieve large currents. By contrast, current-potential curve is steep, and only small values of the current-potential relationship for the cathodic reaction is shallow, because io is small, and the anodic and cathodic branches are not symmetrical (i.e., βa ≠ βc). Because both reactions are occurring on different sites on the same surface,Fig. 1, icorr cannot be measured by coupling the metal to an ammeter. However, Ecorr can be measured against a suitable reference electrode by using a voltmeter with an input impedance high enough to draw insignificant current in the measuring circuit. Figure 8 shows that the value of Ecorr is determined by the shape of the current-potential relationships for the two reactions; that is, it is a parameter with kinetic but not thermodynamic significance. Because its value is determined by the properties of more than one reaction, the corrosion potential is often termed a mixed potential. In the literature, diagrams such as the one in Fig. 8 are often plotted in the form log i versus E, and the algebraic sign of the cathodic current is neglected, so that the anodic and cathodic currents can be plotted in the same quadrant (Fig. 9). Such diagrams are termed Evans diagrams. The two linear portions in an Evans diagram are

the Tafel regions, with slopes given by Eq 30 and the equivalent plot for the cathodic reaction. Sometimes, the nonlinearity close to the equilibrium potentials is ignored, and the curves are plotted as totally linear. This approximation acknowledges that, commonly, the two io values are orders of magnitude lower than icorr and therefore have a negligible effect on the scale of the Evans diagram. It should be noted that the currents plotted in an Evans diagram are the partial currents for the anodic and cathodic reactions, and that the measurable current is the sum of these two partial currents (taking into account that they are opposite in sign). The value of such diagrams is in their use to illustrate the influence of various parameters on the corrosion process.

Fig. 9 The current-potential relationships of Fig. 8 plotted in the form of an Evans diagram. Note: The solid lines are partial anodic and cathodic currents, not measurable currents. Figure 10(a) shows an Evans diagram for the same anodic dissolution process coupled to two different cathodic reactions. Recalling the definition of ΔEtherm from Eq 18, the following can be written: (Eq 39) leading to: (Eq 40) that is, the bigger the difference in equilibrium potentials, the larger the corrosion current. The anodic activation overpotential for the first reaction (

= Ecorr -(Ee)a) is less than that for the second: (Eq 41)

Fig. 10 Evans diagram for one anodic dissolution reaction coupled (separately) to one of two different oxidant reduction reactions. (a) The two oxidant reduction reactions have similar kinetic characteristics (i.e., similar current-potential shapes). (b) The two oxidant reduction reactions have very different kinetic characteristics (i.e., very different current-potential shapes). (c) An anodic dissolution reaction

with a large (icorr)a coupled to an oxidant reduction reaction with a small (icorr)c. The currents labeled C1 and C2 show the effect on log (icorr) and Ecorr of changing the concentration of available oxidant. The value of ΔEtherm is not the only parameter controlling the corrosion rate. Figure 10(b) shows the same situation as in Fig. 10(a) , except that the two cathodic reactions possess very different polarization characteristics (i.e., relationships between current and potential). Despite the fact that

, the

activation overpotential , and the corrosion couple with the largest thermodynamic driving force produces the lowest corrosion current. Inspection of Fig. 10(b) shows this can be attributed to the differences in exchange currents, io, and slopes, hence different Tafel coefficients, βc, for the two cathodic reactions. This situation is common for active metals in acidic or neutral aerated solutions. Even though the thermodynamic driving force for corrosion is greater in neutral solutions containing dissolved oxygen, corrosion proceeds more rapidly in deaerated acidic solutions. This is due to the slowness of the kinetics for oxygen reduction and can be appreciated by comparing the kinetic characteristics for the two processes on iron. Thus,

= 10-3 to 10-2 A/m2, and

≅ 120 mV/decade. By comparison,

≅ 10-10

> 120 mV/decade. A/m2, and Rate Control by the Anodic or Cathodic Reaction. The overall rate of corrosion will be controlled by the kinetically slowest reaction, that is, the one with the smallest exchange current and/or largest Tafel coefficient. This can be appreciated from Fig. 10(c) in which (io)a > (io)c and βa < βc. This leads to a large difference in activation overpotentials, with: (Eq 42) This means the cathodic reaction is strongly polarized and must be driven significantly to achieve the corrosion current. By contrast, the anodic reaction remains close to equilibrium, requiring only a small overpotential (anodic) to achieve the same corrosion current. Under these conditions, the corrosion potential, Ecorr, lies close to the equilibrium potential for the kinetically fastest reaction. If the cathodic reaction was the fastest, then Ecorr → (Ee)C, and the anodic metal dissolution reaction would be rate controlling. If the kinetics of the two half- reactions were similar (i.e., (io)a (io)c and βa βc), then Ecorr would be approximately equidistant between the two equilibrium potentials, and the corrosion reaction would be under mixed anodic/cathodic control. This is the situation illustrated in Fig. 9. The corrosion of iron in aerated neutral solution can be used to illustrate the point. For the metal dissolution reaction, ≅ 10-4 to 10-5 A/m2 and

≅ 50 to 80 mV/decade, whereas for O2 reduction,

≅ 10-10 A/m2 and > 120 mV/decade. Consequently, O2 reduction should be rate controlling, and Ecorr would lie close to the equilibrium potential for iron dissolution. Figure 10(c) also shows the effects of changing the kinetics of the two reactions. Such changes could be caused by increasing the available concentration of cathodic reagent or by increasing the surface area of available metal. Changing the reagent concentration would also lead to a change in (Ee)c, but this is ignored in Fig. 10(c) for the sake of clarity. Changes in the kinetics of the fast anodic reaction are reflected in changes in the value of Ecorr (Δ(Ecorr)a large) but have little effect on icorr (Δlog(icorr)a small). However, changes in the kinetics of the slow cathodic reaction have little effect on Ecorr (Δ(Ecorr)c small) but a significant influence on icorr (Δlog(icorr)c large). Thus, in this example, the cathodic reaction is the rate-controlling reaction, and the anodic reaction is said to be potential determining. Such changes in Ecorr can sometimes be used as diagnostic tests for ascertaining the rate-determining step, and the maximum benefit in slowing corrosion can be gained by attending to the rate-determining step. Additionally, a measurement of Ecorr and its evolution with time provides a simple but only qualitative way of tracking the evolution of a corrosion process with time.

Reference cited in this section 6. A.J. Bard and L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications, 2nd ed., John Wiley and Sons, 2001.


Mass Transport Control The relationships between current and potential described previously are valid when the overall corrosion process is under activation control, that is, controlled by the interfacial rate of one or the other of the halfreactions (step 1 or 2 in Fig. 5). However, if the cathodic reagent at the corrosion site is in short supply, then mass transport of this reagent could become rate controlling (Eq 3, Fig. 5). Under these conditions, the cathodic charge-transfer process is fast enough to reduce the concentration of the cathodic reagent at the corrosion site to a value less than that in the bulk solution. Because the rate of the cathodic reaction is proportional to the surface concentration of reagent, the reaction rate will be limited (polarized) by this drop in concentration. For a sufficiently fast charge transfer, the surface concentration will fall to 0, and the corrosion process will be totally controlled by flux of oxidant to the corroding surface. In the case of one-dimensional transport, this flux (J) can be calculated from the solution of Fick's first law: (Eq 43) at the corroding site (x = 0), where D is the diffusion coefficient, and cO is the reagent concentration at point x. Diffusion gradients evolve with time, according to Fick's second law: (Eq 44) and, for non-steady-state conditions (i.e., ∂c/∂x at x varies with time), conditions are complicated. General corrosion processes, however, generally occur under steady-state conditions, because convective flow of the environment occurs (e.g., flow down a pipe). For this situation, a simple analysis can be achieved by linearizing the concentration-distance profile according to the Nernst diffusion layer treatment (Fig. 11). This linearization yields a demarcation line at a distance, δ, from the surface such that, for x > δ, the bulk concentration is maintained by convection. For x < δ, the oxidant is assumed to be transported to the surface by diffusion only. This solution layer is termed the diffusion layer, and its thickness is determined by the local geometry at the site and the solution velocity (degree of convection). Of course, this is an approximation, because the transition from convective flow to diffusive transport is more gradual and not so strictly demarcated.

Fig. 11 Illustration showing the concentration-distance profile for oxidant O involved in a corrosion process with a metal in which the overall process is partially controlled by convective-diffusive transport of the oxidant to the corroding surface. The dashed-dotted lines show the concentration profile assumed in the Nernst diffusion layer treatment. Using this simplified treatment, Eq 43 can be written: (Eq 45) where is the oxidant concentration at the corroding surface (x = 0), and is the concentration for x ≥ δ. For steady-state conditions, all the oxidant transported down the concentration gradient to the corroding surface must react electrochemically. If it did not, then the concentration at the corroding surface would adjust until it did. This condition of mass balance means that the cathodic current (expressed in A/cm2) must be directly proportional to the flux (expressed in mol/cm2/s). The proportionality constant is given by Faraday's law (Eq 8) and can be written: (Eq 46) The corrosion current is equal to the cathodic current, because the overall corrosion reaction is controlled by the transport of oxidant to the surface; that is, corrosion can only progress at the rate that oxidant becomes available. Under limiting conditions, → 0, and a maximum transport-controlled corrosion current is obtained: (Eq 47) When corrosion is occurring at this limiting rate, the rate can only be increased or decreased by varying either the bulk concentration of oxidant, , or the thickness of the diffusion layer, δ. This limitation on corrosion rate is termed concentration polarization and is illustrated in the Evans diagram in Fig. 12. For a small shift of the potential away from the equilibrium potential (point 1), = , and there is no limitation on reagent supply. The current remains in the Tafel region; that is, charge transfer is still rate controlling, and the overpotential is purely an activation overpotential (

). For a larger shift of potential from

(Ee)c, < (point 2), and the current is less than expected on the basis of activation control; that is, the current follows the solid line as opposed to the dashed-dotted line. Under these conditions, the current is partly

activation and partly transport controlled, and the total overpotential (ηT) is the sum of an activation (ηA) and a concentration (ηC) overpotential: ηT = ηA + ηC

(Eq 48)

For a sufficiently large polarization from equilibrium, → 0, the current becomes independent of potential, and the cathodic current is now at the maximum given by Eq 47.

Fig. 12 Partial Evans diagram (i.e., showing only the partial current for the oxidant reduction reaction) for mixed activation-concentration polarization The effect of a number of corrosion parameters for a corrosion process proceeding under various degrees of activation/mass transport control can be assessed using an Evans diagram, as shown in Fig. 13. Three situations are considered. For cathodic curve 1, corrosion occurs with the anodic reaction totally activation controlled ( = ) and the cathodic reaction totally mass transport controlled ( = ), that is, = 0. If the solution is now made to flow, the thickness of the diffusion layer (δ) will decrease, (ic)lim (Eq 47) will increase and hence so will the corrosion current. The figure shows that Ecorr shifts to a more positive value, and: (Eq 49) decreases due to a decrease in

. For more vigorous stirring,

corroding surface is now sufficiently large to maintain

=

reaches 0, because the flux of oxidant to the . Again, Ecorr shifts to a more positive value to

because = 0. The cathodic reaction is now fully activation controlled ( = ), reflect this decrease in and further increases in fluid velocity will not affect the corrosion rate. Such changes in Ecorr and icorr with the degree of convection can be used to indicate whether mass transport control is operative. If the anodic, as opposed to the cathodic, reaction was mass transport controlled, then a similar analysis would apply, but Ecorr would shift to more negative values with increasing degree of convection.

Fig. 13 Evans diagram showing the influence on Ecorr and log (icorr) of changing the transport rate of oxidant O to the corroding surface More generally, the term DO/δ is termed the mass transport coefficient (mc), and Eq 46 is written in the form: (Eq 50) An equation for the corrosion rate under activation control can also be written in the simplified form: (Eq 51) where kc can be considered a potential-dependent rate constant for the cathodic reaction. Equation 51 could be considered a restatement of the Butler-Volmer equation, and the reader is referred to standard electrochemical text books for more detailed descriptions (Ref 6). Combining Eq 50 and 51 and eliminating yields: (Eq 52) where kc can be considered the activation control parameter and mc the mass transport control parameter. Whether activation or mass transport kinetics determine the corrosion rate is straight forwardly appreciated by considering the relative values of mc and kc in Eq 52. For mc » kc, the bracketed term reduces to kc, and the corrosion current is activation controlled. For mc « kc, the term reduces to mc, and corrosion would be mass transport controlled. While Eq 52 may define the relative importance of activation and mass transport control, it does not contain any information on the factors that control mass transport and hence determine the value of mc. The dependence of mc on solution flow rate can be determined experimentally, and its form varies, depending on the geometry of the system. In general, this dependence takes the form: icorr α fn

(Eq 53)

where f is the flow rate, and n is a constant that depends primarily on the geometry of the system. For flow over a flat plate, n is 0.33 for laminar flow (smooth, Re < 2200), where Re is the Reynold's number (defined subsequently). The variation of mc depends not only on flow rate but also on properties such as the kinematic viscosity of the fluid (ν), the diffusion coefficient of the species (D), as well as the geometry of the system. These effects can be specified using dimensionless parameters such as the Reynold's; Re; and Schmidt, Sc, numbers given by: (Eq 54)

(Eq 55) The Reynolds number is a measure of the ratio of convective to viscous forces in the fluid, and, for laminar flow down a pipe, L is a characteristic length. The Schmidt number quantifies the relationship between hydrodynamic and diffusion boundary layers at the corroding surface. Thus, it can be shown that for flow over a smooth, flat, corroding surface: (Eq 56) where α and β are numerical constants that depend on the geometry of the system and the convective flow conditions. Smooth, laminar flow can only be achieved up to a certain Reynold's number or flow rate, f, if L and ν are constant, beyond which the flow becomes turbulent, and the dependence of icorr on flow rate increases. For still higher rates, activation control can be achieved (mc » kc), and the corrosion rate becomes independent of flow rate. This behavior is illustrated in Fig. 14. The solid line shows the effect of flow rate when the corrosion reaction is fast (kc large), and a large flow rate is required to achieve activation control (mc » kc). The dotted line shows the behavior expected for a slow corrosion reaction (kc small), when only a small flow rate is required to achieve activation control.

Fig. 14 The influence of solution flow rate on the corrosion rate (expressed as a current) showing the response in different flow regimes

Reference cited in this section 6. A.J. Bard and L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications, 2nd ed., John Wiley and Sons, 2001.


Influence of Corrosion Product Deposits

Thus far in the discussion of transport effects, only the transport of the cathodic reagent (step 3 in Fig. 5) has been considered. The transport of the anodic (metal) dissolution product (step 4 in Fig. 5) also affects the corrosion rate but in a different way. If dissolved metal ions are allowed to build up at the corroding surface, supersaturation of solid oxides/hydroxides can occur, leading to film formation reactions (step 5 in Fig. 5). These corrosion product deposits will form at a rate determined by the rate of the corrosion process and the solubility of the metal cations in the particular exposure environment. Determining their influence on the corrosion rate is not simple. A key feature is the porosity of the deposit (ε), because this determines the area of metal left exposed and also reduces the diffusion coefficient of both the anodic dissolution product and the cathodic reagent by a factor directly proportional to the porosity. The pores are also likely to be nonlinear, and their effective length will be greater than the thickness of the deposit. The diffusion coefficient will, therefore, also be attenuated by a tortuosity (τ) factor. The effective diffusion coefficient (Deff) will be given by: Deff = ετDOx

(Eq 57)

This situation is illustrated schematically in Fig. 15, and Eq 45 for the corrosion rate could be rewritten to yield: (Eq 58) where corrosion is only occurring on a fraction (εA) of the original exposed metal area, A, and l is the thickness of the corrosion product deposit. It should be noted that this represents a very simplified discussion of what could be a much more complicated process.

Fig. 15 Schematic illustrating how the presence of a corrosion product deposit influences the corrosion of an underlying metal by limiting both the area of exposed metal and the diffusion of oxidant to the corroding surface. D, diffusion coefficient; ε, porosity; τ, tortuosity factor; A, area


Kinetics of Corrosion of Passive Metals The formation of a passive film (step 7 in Fig. 5) can have a large influence on the corrosion rate. Passivation becomes thermodynamically possible when the corrosion potential exceeds (becomes more positive than) the potential corresponding to the equilibrium between the metal and one of its oxides/hydroxides: Ecorr > (Ee)M/MO

(Eq 59)

Inspection of the Pourbaix diagram for the metal/ metal oxide/aqueous solution system shows that this condition moves Ecorr into the oxide stability region, as illustrated in Fig. 16 for the iron/iron oxide/water system. For point 1, Ecorr < (Ee)M/MO, and corrosion of the bare metal is expected, while, for point 2, Ecorr > (Ee)M/MO, and the metal should be oxide covered and passive. It should be noted that experimentally measured primary passive potentials inevitably do not correspond directly with these thermodynamically determined potentials, because kinetic as well as thermodynamic effects are involved in oxide film formation.

Fig. 16 Simplified Pourbaix (equilibrium potential-pH) diagram for the iron-water system. The upper dashed line shows the potential for H2O in equilibrium with O2. The current-potential or polarization curve for the anodic process is shown in Fig. 17 and can be divided into a number of regions. In region AB, the active region, metal dissolution is unimpeded, because no passive film is present. Because such polarization curves are usually recorded under activation control conditions, local supersaturation leading to the deposition of corrosion products (discussed previously) is generally assumed not to influence the active dissolution of the metal. However, for real corrosion conditions, and especially for reactive metals such as iron, the current in this region could be suppressed by corrosion product deposits, as discussed previously. For the purposes of the present discussion, it is assumed that this does not occur. For active dissolution, therefore, the current should conform to the Tafel relationship (Eq 30), and its extrapolation back to (Ee)a would yield a value of (io)a.

Fig. 17 Polarization curve for a metal that undergoes active-to-passive and passive-to-transpassive transitions At a potential B, shown in Fig. 17 to coincide with (Ee)M/MO (although, as noted previously, this is rarely the case experimentally), there is a departure from the Tafel relationship, leading eventually to a decrease in current to a low value. The electrode is said to have undergone an active-to-passive transition and, by point C, has become passive. The potential at point B may or may not correspond to the potential (Ee)M/MO. Thermodynamics demands only that the condition in Eq 59 be satisfied. The maximum current achieved immediately before the transition is termed the critical passivating current (icrit). The potential at which the current falls to the passive value is called the passivation potential (Epass), and it corresponds to the onset of full passivity. For gold, silver, and platinum, the passivation potential is close to (Ee)M/MO, but for most other metals, it is much more positive than this equilibrium value. For E > Epass, the metal is said to be in the passive state, and the current is very low. A comparison of critical currents to passive currents puts the industrial importance of passivity into perspective. Critical passivating currents (expressed as current densities) can be as high as tens of mA/cm2, while passive currents can be as low as nA/cm2. Consequently, from the industrial corrosion perspective, the establishment of passivity is an important feature. A detailed discussion of the properties of passive films is beyond the scope of this article (see the article “Passivity” in this Volume). A number of theories exist to explain the growth kinetics and properties of oxide films, but the discussion of one of them, the point defect model (Ref 7), serves to illustrate the critical features of the current-potential behavior in the passive region. The essential features of this model are shown schematically in Fig. 18. The film is assumed to be composed of an inner barrier layer, whose insulating properties provide the essential corrosion protection, and an outer recrystallized layer that confers little extra corrosion protection and is composed of hydrated metal species. This outer layer may not exist in acidic solutions in which the metal cations are highly soluble (Fig. 6). Considerable experimental evidence exists to support this structure.

Fig. 18 Schematic description of the point defect model for the growth of a passive oxide film Oxide growth is assumed to occur by the transport of defects through the film under the influence of the electric field that exists within the oxide. The nature of the defect depends on the metal/alloy and the nature of the growing oxide. However, the key vacancies are anion vacancies created at the metal-oxide interface, cation vacancies (Vm) created at the oxide- solution interface, and cation interstitials created at the metal-oxide interface. The general observation that the current in the passive region is independent of potential can be interpreted in terms of this model. As the potential is increased through the passive region, a progressive thickening of the oxide occurs, such that the electric field within the oxide remains constant. Thus, the steadystate current at each potential is a balance between oxide formation at the metal-oxide interface and

dissolution/recrystallization (to the outer hydrated layer) at the oxide-solution interface. Reference 7 provides more extensive discussions. For high anodic potentials, the observed behavior depends on the nature of the metal and the properties of the oxide film. For metals that form insulating oxides (e.g., aluminum, titanium, zirconium), the films continue to thicken with increasing potential, and low passive currents are maintained. For other metals, for example, iron, chromium, nickel, and materials such as stainless steels and nickel-chromium alloys, oxygen evolution can occur on the outside of the passive film once the potential exceeds point E (Fig. 17): 2H2O → O2 +4H+ + 4e-

(Eq 60)

For this reaction to occur, the film must be electronically conducting. This is possible because the passive films formed are thin (nanometers) and possess semiconducting properties. The dashed-dotted line in Fig. 17, in the potential region D to E, corresponds to the phenomenon of transpassivity. In this region, the oxide film begins to dissolve oxidatively, generally as a hydrolyzed cation in a higher oxidation state than that which exists in the film. An example would be the dissolution of chromium, present in the passive film on stainless steels as Cr III, as CrVI in the form of While transpassive dissolution, starting at point D, and O2 evolution, starting at point E, are shown as clearly separated in Fig. 17, they often occur together, and a distinct transpassive region cannot be observed. When coupled to a cathodic reaction in a corrosion process, a number of criteria must be satisfied if the metal is to be in the passive region and, on normal industrial time scales, free of corrosion: • •

The equilibrium potential for the cathodic reaction ((Ee)c) must be greater than the passivation potential (Epass). The cathodic reaction must be capable of driving the anodic reaction to a current in excess of the critical passivation current, icrit, in Fig. 17.

Three possible situations are shown in the Evans diagram in Fig. 19. The solid line shows the anodic polarization curve, and lines 1, 2, and 3 show the cathodic polarization curves for three different cathodic halfreactions (On + ne- → Rn).

Fig. 19 Effect of various cathodic reactions on the corrosion current and potential for a metal capable of undergoing an active-passive transition For cathodic reaction 1 (Fig. 19), (Ee)c1 < Epass. Because the corrosion potential must lie between (Ee)a and (Ee)c1 for the two reactions to form a corrosion couple, the required condition for passivation, Ecorr > Epass, cannot be achieved. Therefore, Ecorr remains in the active region, and the metal will actively corrode.

For cathodic reaction 2 (Fig. 19), the condition (Ee)c2 > Epass is met, but the two polarization curves intersect at an anodic current Epass and i > icrit are both met. As a consequence, Ecorr > Epass, and the metal passivates, with the corrosion current equal to the passive dissolution current. It is clear from this discussion that mild oxidizing agents (ΔEtherm = (Ee)c - (Ee)a small) could leave the material susceptible to corrosion, and strong oxidizing agents (ΔEtherm large) are required to ensure the metal will be in the passive region. However, in the presence of extremely strong oxidizing agents, it is possible for Ecorr to become sufficiently positive that physical instabilities in the oxide become feasible. These minor film breakdown events can lead to many forms of localized corrosion if allowed to stabilize. A discussion of these processes is beyond the scope of this article, but the danger in fracturing the passive film can be appreciated from a consideration of the extrapolation of the anodic active region, shown by the dashed-dotted line in Fig. 19. If the oxide did not form, then this extrapolation shows that currents at positive potentials would be very large. Thus, breakdown of the passive film at a local site within the passive region would lead to an extremely high local current density and the possibility of deep localized corrosion of the metal.

Reference cited in this section 7. D.D. Macdonald, Pure Appl. Chem., Vol 71, 1999, p 951


References 1. J.H. West, Electrodeposition and Corrosion Processes, Van Nostrand Reinhold, 1971 2. D. Jones, Principles and Prevention of Corrosion, 2nd ed., Prentice Hall, 1996 3. L.L. Shreir, R.A. Jarman, and G.T. Burstein, Ed., Metal/Environment Interactions, Corrosion, Vol 1, 3rd ed., Butterworth/Heinemann, 1994 4. L.L. Shreir, R.A. Jarman, and G.T. Burstein, Ed., Corrosion Control, Corrosion, Vol 2, 3rd ed., Butterworth/Heinemann, 1994 5. C.F. Baes and R.E. Mesmer, The Hydrolysis of Cations, John Wiley and Sons, 1976. 6. A.J. Bard and L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications, 2nd ed., John Wiley and Sons, 2001. 7. D.D. Macdonald, Pure Appl. Chem., Vol 71, 1999, p 951

F. Huet, R.P. Nogueira, and H. Takenouti, Aqueous Corrosion Reaction Mechanisms, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 52–60

Aqueous Corrosion Reaction Mechanisms François Huet, Ricardo P. Nogueira, Bernard Normand, and Hisasi Takenouti, Université Pierre et Marie Curie and UPR 15 of CNRS, “Laboratoire Interfaces et Systèmes Electrochimiques,” Université Pierre et Marie Curie

Introduction CORROSION OF METALLIC MATERIALS is governed by electrochemical kinetics, so that the general concepts developed for studying electrochemical reaction mechanisms may be applied to corrosion. The processes of charge transfer taking place at the electrode interface within the double layer and of mass transport at the vicinity of the electrode surface are introduced first. Then, the corrosion processes, which involve anodic and cathodic reactions at specific electrode sites, are described briefly. Some reaction mechanisms for cathodic and anodic processes are presented to illustrate the great variety of reaction mechanisms occurring at the electrode interface. Some experimental methods for devising a reliable reaction model are also described. The presentations given here are limited to a survey of each item; readers wishing more detailed information are invited to consult the many books devoted to electrode kinetics and corrosion processes (Ref 1, 2, 3, 4).

References cited in this section 1. H. Kaesche, Metallic Corrosion, Principles of Physical Chemistry and Current Problems, NACE International, Houston, TX, 1985 2. F. Mansfeld and U. Bertocci, Ed., Electrochemical Corrosion Testing, STP 727, ASTM International, 1981 3. J.C. Scully, Ed., Treatise on Materials Science and Technology, Vol 23, Corrosion: Aqueous Processes and Passive Films, Academic Press, 1983 4. J.R. Scully, D.C. Silvermann, and M.W. Kendig, Ed., Electrochemical Impedance: Analysis and Interpretation, STP 1188, ASTM International, 1994

F. Huet, R.P. Nogueira, and H. Takenouti, Aqueous Corrosion Reaction Mechanisms, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 52–60 Aqueous Corrosion Reaction Mechanisms François Huet, Ricardo P. Nogueira, Bernard Normand, and Hisasi Takenouti, Université Pierre et Marie Curie and UPR 15 of CNRS, “Laboratoire Interfaces et Systèmes Electrochimiques,” Université Pierre et Marie Curie

Fundamental Aspects of Electrode Kinetics Electrochemical reactions may be considered as physical and chemical processes of electric charge transfer between two phases containing different charge carriers: electrons in the metallic phase and ions in the electrolyte. This charge transfer induces an electric current flow and chemical transformations at the interface.

A thorough investigation of an electrochemical process involves the identification of reactants, the separation of elementary interfacial reactions, and the establishment of a reaction mechanism. Analytical methods are generally applied to identify the initial, final, and, sometimes, intermediate species, while kinetic methods are used to identify the links between the elementary reaction steps and assess their rate constant. Kinetic methods differ from other methods by the following points: • • •

Reactions are localized at the interface so that they concern a tiny volume and a small number of identifiable entities. Most in situ surface analytical techniques cannot be applied because of the presence of electrolyte. The kinetics are easily characterized by electric signals (the current flowing across the interface and the electrode potential) with more and more sophisticated methods.

The rate of charge transfer reactions governed only by the activation energy is derived first. Then, the case in which mass transport modifies the concentration gradient of the species at the interface and, therefore, changes the overall reaction rate is discussed.

Charge-Transfer Process Charge transfer may take place as a single step determined by the activation energy or as a multistep transfer involving intermediate species in the reaction. Single-Step Charge Transfer Reaction Rate. Different macroscopic descriptions allowed the absolute reaction rate, k, to be associated with the change in activation energy (Gibbs free energy, or ΔG‡), through the Arrhenius equation: (Eq 1) where C is a constant, T is the absolute temperature (in K), and R is the gas constant (8.31 J/mol · K). The activation energy change ΔG‡, in electrochemical reactions is not the difference in free energy, ΔG, between the initial and final states of the reactants; it takes into account the height of the energy barrier, as will be explained, and depends on the cell potential E. For a reaction involving the forward and backward single-step transfer of z electrons per molecule of the species considered, the simplest situation, which leads to the empirical Tafel law between the current density and the electrode potential, consists in postulating that a change in potential (ΔE) induces a change in free energy, ΔG‡, according to the following relationship. For the anodic (oxidation) reaction: (Eq 2) For the cathodic (reduction) reaction: (Eq 3) where F is the Faraday constant (96,485 C/mol), α is the charge-transfer coefficient (0 ≤ α ≤ 1). The subscript 0 indicates that all reactants are in a standard state (constant temperature, pressure, potential, activity, etc.). Equations 2 and 3 indicate how the electric energy shift zFΔE of the z moles of electrons induced by the electrode potential change (ΔE) is shared between the two partial reactions. This may be illustrated in Fig. 1 that shows the free energy evolution during both the reduction (from left to right) and oxidation reactions (from right to left). The solid lines correspond to the standard conditions for both reactions; in that case (ΔE = 0), Δ

represents the energy required by one mole of oxidized species to reach the activated state, A, from the

initial stable state (minimum energy), and be reduced. Δ represents the energy required by one mole of reduced species to reach the activated state, A, from the initial stable state (minimum energy) and be oxidized. The dashed line corresponds to the activation energy of one mole of oxidized species and z moles of electrons during the reduction reaction after a potential change ΔE. The anodic curve was arbitrarily kept unchanged.

Fig. 1 Schematic representation of the influence of a potential change on the activation energy during charge transfer in an oxidation-reduction reaction It can be seen graphically that if α represents the fraction of the electric energy zFΔE devoted to the activation energy of the oxidation reaction, the height of the energy barrier is decreased by a factor αzFΔE for the oxidation reaction (Eq 2) while it is increased by a factor Δ −Δ = zFΔE - αzFΔE for the reduction reaction (Eq 3). The dependence of the anodic and cathodic rate constants, ka and kc, on the electrode potential is obtained as a consequence of Eq 1, Eq 2, Eq 3:

(Eq 4)

Current-Potential Relationship. The current provided by the oxidation-reduction reaction is proportional to the net production rate of the oxidized species:

(Eq 5)

where i is the current density, and , (in mol/cm3) are the concentrations of the oxidized and reduced species at the electrode interface. For a purely charge-transfer controlled reaction, the mass-transport processes are fast so that the concentrations at the interface are equal to the concentrations in the electrolyte bulk and do not depend on the electrode potential. It is convenient to introduce the standard equilibrium potential, E0, which corresponds to i = 0 for identical concentrations of the oxidized and reduced species in the solution. In that case, since the reaction rates k0,a and k0,c are equal (= k0, termed exchange rate constant of the reaction, in cm/s), i can be expressed as:

(Eq 6)

This equation gives the expression (Nernst law) of the equilibrium potential (Eeq) in the solution considered (

≠

):

(Eq 7) from which the overpotential, η = E - Eeq, is defined. Equation 6 gives the Butler-Volmer equation, which is the basis of the electrochemical kinetics (Ref 5, 6): (Eq 8) where the exchange current density (i0) is defined as: (Eq 9) gives information on the degree of reversibility of the reaction. The Butler-Volmer equation gives the partial anodic and cathodic currents. At sufficiently high anodic overpotentials, the cathodic process becomes negligible. At sufficiently high cathodic overpotentials, the anodic process becomes negligible. The anodic ia and cathodic ic current densities are then:

(Eq 10)

According to the International Union of Pure and Applied Chemistry (IUPAC) recommendation, the cathodic current and its Tafel coefficient, -(1 - α)zF/RT, are expressed by negative values. In these equations (Tafel laws), the logarithm of the current density (in absolute value) is a linear function of the electrode potential. Figure 2 summarizes the current behavior given by Eq 8 and 10.

Fig. 2 Butler-Volmer electrode kinetics. i0 = 10 mA/cm2, charge transfer coefficient, α = 0.75, charge number, z = 1, T = 298 K Multistep Charge Transfer. In the single-step charge transfer, z electrons are exchanged simultaneously. However, this requires a much higher activation energy than consecutive elementary reaction steps involving only one electron transfer. As a simple example, consider the situation where two electrons are transferred (z = 2) via two irreversible reactions involving an adsorbed intermediate species, Bad. (Eq 11) (Eq 12) M is the metal and Asol and Csol are species in solution. For example, the dissolution of iron in sulfuric acid at low current may be modeled by these reactions with M, Asol, Bad, and Csol representing, respectively, Fe, OH-, (FeOH)ad, and Fe2+ in solution (Ref 7). Bad covers a fraction (θ) of the electrode surface area (0 < θ < 1) so that the reaction in Eq 11 takes place only at the fractional area of 1 - θ. If a Langmuir-type adsorption isotherm can be assumed, the superficial concentration (in mol/cm2) (CB) of Bad is equal to βθ, where β is the maximum concentration of Bad (in mol/cm2), while the concentration CM of atoms M available for the reaction shown in Eq 11 is β(1 - θ). For first-order kinetic reactions, the mass and charge balances can be expressed, respectively, as follows when the concentration of Csol at the interface can be assumed to be constant: (Eq 13)

(Eq 14) where b1 and b2 are the Tafel coefficients and k0,1 and k0,2 are the reaction rates, expressed in s-1 at the potential = k0,iβ in mol/cm2 per E = 0 V, defined with an arbitrary potential reference. In the literature, reaction rates s are frequently introduced so that the parameter β appears in Eq 13 only. Often, β values of about 2 to 5 × 10-9 mol/cm2 are used; they correspond to one adsorbed molecule per one surface atom for an electrode with a certain surface roughness. At the steady state (subscript s), there is no time change in the potential, current, and surface coverage; the righthand side of Eq 13 is then equal to zero, which gives the steady-state value of θ: (Eq 15) The equation of the steady-state polarization curve follows, from Eq 14: (Eq 16) If b1 > b2, the current density is determined by the reaction in Eq 11 at low potentials (k1 « k2) and by the reaction in Eq 12 at high potentials (k1 » k2) leading to a Tafel plot [log(i) - E] with two different slopes, b1 at low potentials and b2 at high potentials. The surface concentration of the adsorbed intermediate species β · θ or its potential dependence cannot in general be observed experimentally by steady-state methods. In contrast, the impedance technique or other time or frequency resolved methods can track the relaxation of θ induced by a potential change.

Mass Transport Process and Mixed Kinetics The Butler-Volmer equation is valid for systems under pure charge transfer control; the mass transport is much faster than the charge transfer and ensures the constant supply of reactants at the interface. This is not the case

for fast charge- transfer reactions and differences in the concentration of the reacting species, and the interface, and Cox and Cred, in the bulk of the electrolyte appear. Equation 8 is then rewritten as:

, at

(Eq 17)

In this situation, the current density is determined by both the diffusion and activation processes (mixed kinetics). Under the assumption of linear concentrations in the Nernst diffusion layer of thickness δ, the mixed current density (imix) is: (Eq 18) where Dred is the diffusion coefficient (cm2/s) of the reduced species, since the flux of these species consumed in the reaction is balanced by the flux arriving by diffusion. the full expression of imix:

can be calculated with Fick's law, which gives

(Eq 19) where the pure activation current density (iact) is given by Eq 6: iact = zF(kaCred - kcCox)

(Eq 20)

Mass transport of both oxidized and reduced species is supposed here to be controlled by diffusion; if the reaction rate is controlled by diffusion of only one species, the diffusion coefficient of the other species is taken to be infinite. If the reaction rate is fast enough, the concentration of the species consumed at the interface tends to be zero. The overall reaction is under mass-transport control, and imix tends to be the diffusion limiting current density (ilim) (given by Eq 18 with = 0), which is independent of the potential. For slower kinetics, the current densities imix, iact, and ilim are related by the following equation when the reverse reaction is negligible: (Eq 21) Figure 3 presents the polarization curves when the overall reaction rate is influenced by the diffusion process. The kinetic constants determining the pure activation current density iact are those used in Fig. 2, while the diffusion limiting current densities are set to 25 mA/cm2 for both anodic and cathodic reactions. The deviation between the current-density curves (imix,a and imix,c) for the mixed kinetics and iact,a and iact,c for the pure activation-controlled kinetics can be noted. It should be emphasized that the diffusion effects are already present at the equilibrium potential since the exchange current density (

) is lower in Fig. 3 than i0 in Fig. 2.

Fig. 3 Polarization curve for mixed kinetics. Same rate constants as for the activation process in Fig. 2. ilim = 25 mA/cm2 for both reactions

References cited in this section 5. K.J. Vetter, Electrochemical Kinetics, Academic Press, 1967 6. A.J. Bard and L.R. Faulkner, Electrochemical Methods, Fundamentals and Applications, John Wiley & Sons, Inc., 1980 7. I. Epelboin and M. Keddam, J. Electrochem. Soc., Vol 117, 1970, p 1052


Electrode Kinetics near the Corrosion Potential The specifics of electrode kinetics near the zero-overall current are described with respect to the reversible process presented earlier. It is shown that these kinetics may be advantageously analyzed by dynamic techniques such as electrochemical impedance spectroscopy. Coupling of Anodic and Cathodic Processes. Though corrosion is basically an electrochemical process, no overall current flow is observed during the free corrosion of a metal (M) at the open-circuit potential, because the anodic process that produces electrons and the cathodic process that consumes electrons, which can be written formally: (Eq 22) (Eq 23) are taking place simultaneously and at the same rate. A more detailed description of these processes is shown later in this article. The reactions in Eq 22 and 23 are not the forward and backward branches of the same reversible reaction, so that they cannot be referred to the same equilibrium potential. The simplest way to describe the kinetics of both processes is to consider that each reaction obeys the Tafel law (Eq 10), which is often experimentally verified because of the large difference in the equilibrium potentials of the anodic and cathodic processes. The potential must be expressed with respect to the same arbitrarily chosen reference electrode, which gives for the anodic and cathodic current density, respectively, from Eq 6: (Eq 24) As stated earlier, free corrosion of metals provides a zero overall current, so that the following relationship at the corrosion potential (Ecorr) allows the corrosion rate icorr to be derived: (Eq 25) The current density near the corrosion potential is then:

(Eq 26) This equation is very similar to Eq. 8, except that the Tafel coefficients ba and bc are no longer mutually related by the charge-transfer coefficients (α and 1 - α), and that Ecorr is a mixed potential and not a thermodynamic equilibrium potential. Differentiation of Eq. 26 yields the slope of the polarization curve at the corrosion potential, which is the reciprocal of the polarization resistance (Rp): (Eq 27) In this equation, known as the Stern-Geary equation, Rp is given as a function of the Tafel coefficients ba(>0) and bc( 0) or a capacitive loop (ImZF < 0) appears in the Nyquist plot of ZF in the complex plane (ReZ, -ImZ) as summarized as follows: ∂i/∂θ > 0 (catalyzer) dθ/dE > 0 (adsorption) dθ/dE < 0 (desorption) Inductive loop Capacitive loop ∂i/∂θ < 0 (inhibitor/passivating species) Capacitive loop Inductive loop Thus, the adsorption of a catalyzer on the electrode surface and, alternatively, the desorption of an inhibitor lead to an inductive loop in the faradaic impedance. Conversely, the passivation process induces a capacitive behavior. Equation 32 shows that ZF is reduced to the charge-transfer resistance (Rt) at high frequency since the term (1 + jωτ) tends to infinity. Moreover, from Eq 29, it can be seen that 1/Rt = icorr(ba - bc) since dθa/dE and dθc/dE vanish at high frequency where di/dE is the inverse of the charge-transfer resistance. In other words, Rt fulfills the Stern-Geary equation when the electrode surface is divided into numerous anodic and cathodic local cells with potential-dependent fractional area, while Rp does not. Thus:

(Eq 33) More generally, if several parameters (p) with different associated time constants are involved in the reaction process, several capacitive and/or inductive loops will appear in the impedance diagram. It is worth remembering that the diffusion impedance cannot be represented by a well-defined time constant (τ) as that presented in Eq 32; it shows a more complicated behavior. However, Eq 33 remains valid even when the electrochemical system is partly or completely limited by mass transport (Ref 9). It is also important to note that ZF cannot be directly measured. Indeed, the impedance of the metal-electrolyte interface, which is accessible to measure, contains the contributions of the electrolyte (solution) resistance (Re) and that of the interface capacitance (double-layer capacitance) Cd, in addition to the faradaic component (ZF). Figure 5(a) presents the electric equivalent circuit of the interface and Fig. 5(b) represents the schematic Nyquist diagram of Z when the product of ∂i/∂θ|E and C in Eq 32 is negative, which gives the capacitive behavior of ZF.

Fig. 5 Impedance models. (a) Equivalent electrical circuit of the interface; (b) Nyquist diagram of Z for a negative product of ∂i/∂θ|E and C in Eq 32

References cited in this section 8. C. Wagner and W. Traud, Z. Elektrochem., Vol 44, 1938, p 391 9. I. Epelboin, C. Gabrielli, M. Keddam, and H. Takenouti, Electrochemical Corrosion Testing, F. Mansfeld and U. Bertocci, Ed., STP 727, ASTM International, 1981, p 150–192


Mechanisms of Cathodic Processes Cathodic processes are half-cell reactions necessary to allow corrosion reactions to take place. Indeed, the metal-dissolving anodic reaction is forcibly coupled to a cathodic counterpart that consumes the electrons produced by the anodic reaction, so that the overall electronic flux is zero for any free-corroding electrode. In principle, any reduction reaction involving species present in the electrolyte and an equilibrium potential which is higher than that of the metal dissolution can act as the cathodic branch of the overall corrosion reaction (see

Pourbaix, Ref 10, and Evans, Ref 11, diagrams for thermodynamic and kinetic information, respectively). In the case of corrosion in an aqueous medium, the main cathodic processes, which are summarized in the following paragraphs, are easy to identify because of the ubiquitous presence of protons and dissolved oxygen that are active charge acceptors.

Proton Reduction The main cathodic process in acidic media is proton reduction, or, more precisely, hydronium ion reduction, according to the overall reaction: 2H+ + 2e- → H2

(Eq 34)

The final reaction product is dissolved molecular hydrogen that can lead to H2-bubble evolution over the metallic surface. The reaction in Eq 34 is the overall representation of a multistep reaction mechanism usually referred as the Volmer-Tafel-Heyrovsky route (Ref 5). The first step (Volmer reaction) is the proton discharge at the interface that yields an adsorbed hydrogen atom, Had, at the metallic surface: H+ + e- → Had

(Eq 35)

Then, Had desorbs in the form of dissolved molecular gas according to two possible pathways. In the VolmerHeyrovsky route, the hydrogen desorption is an electrochemical reaction involving one proton and one adsorbed Had: H+ + Had + e- → H2

(Eq 36)

In contrast, the Volmer-Tafel route is the chemical recombination of two atomic adsorbates without electron transfer: Had + Had → H2

(Eq 37)

The reactions in Eqs 35, 36, 37 are the typical cathodic processes in acidic solutions for most metals (Fe, Zn, Al, etc.) and alloys. Nevertheless, on the basis of the thermodynamic equilibrium analysis (Pourbaix diagram), noble metals, as well as copper, are not expected to corrode with hydrogen evolution, since their dissolution equilibrium potential is higher than that of the proton reduction. However, the presence of complexing agents (for instance, chloride) may act as a reaction catalyst and give rise to copper corrosion with hydrogen evolution.

Oxygen Reduction Aqueous solutions in contact with air contain different amounts of dissolved oxygen depending on various parameters, such as temperature and electrolyte composition. In practice, most corrosion processes take place in such aerated environments so that dissolved oxygen is almost always available for reduction according to the following overall reactions: (Eq 38) (Eq 39) Reactions 38 and 39 represent generic rough simplifications of several possible complex multistep mechanisms that depend on the metal nature and many other parameters. Whatever the specific mechanism, however, aqueous corrosion reactions are often controlled cathodically, which means that the higher the availability of dissolved oxygen in the electrolyte, the higher the corrosion rate. This is why metallic structures in contact with stirred electrolytes or in partially immersed conditions are more prone to severe corrosion attack.

Water Reduction In some cases, when the electrolyte does not have an excess of protons or dissolved oxygen; that is, in neutral or alkaline deaerated media, the cathodic branch of the corrosion reaction may be the reduction of water itself. The overall reaction is: 2H2O + 2e- → H2 + 2OH-

(Eq 40)

which can be split into at least two consecutive steps: water dissociation in H+ and OH- that precedes proton reduction (Eq 34). The reactions in Eqs 34 and 40 are alternative descriptions of the same reaction, which is the obvious consequence of the fact that water is a permanent source of protons. In other words, the most adequate description depends on the pH of the solution. From a thermodynamic point of view, the reaction in Eq 40, as well as the reaction in Eq 34, is not supposed to take place as the cathodic counterpart of the corrosion of metal (M) if the dissolution equilibrium potential of is higher than that of the H+/H reaction. For these reasons, copper is not expected to corrode in acidic media or in deaerated media.

References cited in this section 5. K.J. Vetter, Electrochemical Kinetics, Academic Press, 1967 10. M. Pourbaix, Lectures on Electrochemical Corrosion, Plenum Press, 1973 11. U.R. Evans, The Corrosion and Oxidation of Metals, Arnold Publications, Inc., 1960


Mechanisms of Anodic Processes Mechanisms of metal dissolution have been widely investigated in the literature and recently reviewed by Keddam (Ref 12). Examples are given below for pure metals (iron, copper) and a binary alloy (Fe-17Cr) corroding in acidic solutions. The emphasis of these examples is on the necessity of coupling the classical experimental techniques (polarization curve, electrochemical impedance) with more advanced techniques (ringdisk electrode, electrochemical quartz-crystal microbalance) in order to collect a large amount of kinetic information to make the proposed reaction mechanism as reliable as possible (Ref 13).

Fe Dissolution in H2SO4 The two main dissolution mechanisms reported in the literature, the catalytic mechanism by Heusler (Ref 14) and the consecutive mechanism by Bockris and others (Ref 15) are summarized first. Then, the effect of an organic inhibitor on the dissolution mechanism of iron in an acidic medium is investigated. Catalytic Mechanism. In acidic media, iron atoms dissolve and produce ferrous ions in solution (indicated by the subscript sol) according to the overall dissolution reaction: Fe → (Fe2+)sol + 2e-

(Eq 41)

Heusler proposed the following mechanism for the overall reaction: Fe + OH-

Fe(OH)ad + e-

(Eq 42) (Eq 43)

The adsorbed Fe(OH)ad species formed by the reaction in Eq 42 acts as a catalyzer in the reaction in Eq 43 since it is not consumed. At the steady state, under the assumption that the reaction in Eq 42 is fast (quasi-equilibrium hypothesis), the model predicts a Tafel slope of (1 + 2α2)F/RT in the polarization curve, where α2 stands for the

charge-transfer coefficient in the reaction in Eq 43. Heusler postulated a value of 0.5 for α2; thus this reaction mechanism should exhibit a Tafel slope of 30 mV per decade. This mechanism is often associated with the presence of crystallographic defects on the electrode surface: for example, Fe(OH)ad would be formed at the kink sites, and (OH)ad would move to its neighbor atom when the reaction in Eq 43 proceeds, because the departure of a kink atom creates a new kink. Consecutive Mechanism. According to Bockris and others, the dissolution mechanism differs from the previous one on the second step only. The reaction represented by Eq 43 is replaced by: (Eq 44) (Eq 45) Fe(OH)ad plays the role of a reaction intermediate that is consumed. On the basis of similar hypotheses, the reaction in Eq 42 is fast, and the Tafel coefficient of Eq 44 is equal to α2 = 0.5. The Tafel slope (1 + α2)F/RT of the polarization curve is 40 mV per decade for this mechanism. Hence, under these highly restrictive hypotheses, the mechanism of iron dissolution can be identified by measuring the Tafel slope of the polarization curve. Later, on the basis of EIS data, Keddam and others pointed out that the hypothesis of fast reversible reaction given in Eq 42 was in contradiction with experimental data (Ref 7). They proposed a reaction mechanism similar to that proposed by Bockris and others but in which the reactions given in Eqs 42 and 44 are irreversible. Good agreement was found between the experimental and theoretical time constants of the relaxation of the surface coverage by Fe(OH)ad. The charge-transfer resistance was also in good agreement with that calculated from the Tafel slopes of Eq 42 and 44. These models with a single intermediate species are able to describe the steady state around the corrosion potential. The nonstationary behavior, however, needs more complex models that take into account three or four species such as those proposed later by Keddam and others (Ref 16). Effect of Propargylic Alcohol on the Corrosion Mechanism. Propargylic alcohol (PA) is the simplest molecule in the acetylene group and constitutes one of the main compounds for inhibiting corrosion of iron or carbon steel in acidic medium. The inhibiting effect has been examined by Benzekri and others (Ref 17), who coupled EIS and rotating ring-disk electrode (RRDE) experiments, and revisited by Itagaki and others (Ref 18), who used a channel-flow cell described schematically in Fig. 6. The electrolyte flows from the working electrode (WE) to the detector electrode (DE), which are embedded in an insulating material.

Fig. 6 Schematic representation of the channel-flow cell. Source: Ref 18 At the WE, metal dissolution takes place in parallel with, for instance, hydrogen evolution so that the overall current measured (I) is: (Eq 46) where Idiss is the dissolution current and the hydrogen current. A fraction of the metallic ions produced at the WE is oxidized or reduced on the DE, located downstream, according to the potential applied on the DE that is adequately chosen to avoid oxidizing the dissolved hydrogen produced at the WE. The ratio N0 of the fluxes of ions captured at the DE and produced at the WE, often named collection efficiency, is controlled only

by the cell geometry. It can be determined either by calibration experiments or by theoretical calculations. From the currents measured, I on the WE and ID on the DE, the dissolution current Idiss = ID/N0 and the hydrogen current = I - ID/N0 on the WE can be evaluated. Figure 7 presents the results obtained by Itagaki and others (Ref 18) for an iron WE in sulfuric acid with or without the presence of PA. A DE in gold was used to make the influence of dissolved hydrogen oxidation negligible. The potential of the WE was measured against a saturated (mercurous) sulfate electrode (SSE). Figure 7(a) shows that, up to approximately 10 mA/ cm2, idiss in presence of inhibitor is higher than in its absence. However, this result is not conclusive for the corrosion current since the potential of the electrode was shifted toward more cathodic potentials in the presence of PA that acts essentially to diminish the reduction of hydronium ions. Indeed, weight-loss and impedance measurements have shown that the inhibiting effect of PA essentially decreases the cathodic current at a given potential, so that the overall corrosion process is inhibited (Ref 9).

Fig. 7 Dissolution of Fe in Na2SO4 + H2SO4 (0.5 M , pH 1), with or without addition of 50 mM propargylic alcohol (PA). (a) i: overall working electrode, WE, current density, idiss; Fe dissolution current density. (b) Impedance diagram at 0.5 mA/cm2 (with PA), (c) emission efficiency N(f) = Δidiss/Δi (with PA). Source: Ref 18

According to the dissolution models proposed by Heusler and Bockris and others, the reaction mechanism can be determined from the slope of the polarization curve: 30 mV/decade for the catalytic mechanism (Ref 14) and 40 mV per decade for the consecutive dissolution model (Ref 15). However, the slope of the log(i)-E curve in Fig. 7(a) changed from 11 to 25 mV per decade and that of the log(idiss)-E curve changed from 32 to 48 mV per decade when the inhibitor was added to the solution. None of these slopes correspond to the reaction mechanisms predicted by Heusler (Ref 14) or Bockris and others (Ref 15). The change in Tafel slope would also suggest that the PA was at the origin of the change in dissolution mechanism whereas it acted mainly on the cathodic reaction, as mentioned previously. Figure 7(b) presents the electrochemical impedance measured at the overall current density of 0.5 mA/cm2 in the presence of PA. It is important to mention that the polarization curves in Fig. 7(a) were obtained by a potential sweep method; thus the dynamic behavior investigated in steady-state conditions with the impedance technique was markedly different. The high-frequency capacitive loop, the apex of which was measured at 25 Hz, corresponds to the double- layer capacitance, approximately 80 μF/cm2, in parallel with the charge-transfer resistance, approximately 80 Ω · cm2. Two inductive loops can be seen in the low-frequency range (f < 8 Hz). They were attributed to the relaxation of the surface coverages by a dissolution intermediate species (loop around 5 Hz) and by the adsorbed inhibitor (loop around 1 Hz). The impedance spectra at frequencies higher than 2 Hz and the fact that the product RtI was close to 40 mV suggest that the iron dissolution occurred via the consecutive mechanism proposed by Keddam and others (Ref 7). The fraction of current consumed to form the dissolving species (here Fe2+) is defined at any frequency (f) by the ratio N(f) = ΔIdiss/ΔI derived from the ratio ΔID/ΔI, measured under ac regime, divided by the dynamic collection efficiency. Figure 7(c) shows that N(f) was smaller than 1 above 1 Hz, which indicates that a large part of the ac current was not consumed by the dissolution itself but was used in the formation of adsorbed intermediate species. In contrast, the low-frequency limit of N(f) is close to unity, as expected from the Faraday law. The coupling of advanced experimental techniques offers deeper insight into the reaction mechanism of iron dissolution and also explains how the PA inhibits the corrosion process by decreasing the hydrogen-evolution reaction rate.

Dissolution of Copper in Hydrochloric Acid Copper and copper alloys are widely used in industrial applications because of their high electrical and thermal conductivity and, also, good corrosion resistance in various media. However, the latter decreases significantly in the presence of chloride ions that lead to copper dissolution as cuprous ions at low anodic potentials and cupric ions at higher potentials (Ref 19). Figure 8 presents results of EIS, ring-disk electrode (RDE), and electrochemical quartz crystal microbalance (EQCM) experiments carried out simultaneously on an electrodeposited copper EQCM electrode surrounded by a thin platinum ring (Ref 20). Since the EQCM electrode could not rotate, an impinging jet cell was employed instead of a RRDE to control the hydrodynamic conditions.

Fig. 8 Dissolution of Cu in 1 M NaCl + 0.05 M Na2CO3, pH 8, i = 0.1 mA/cm2. (a) Impedance diagram; (b) N(f) = ΔIdiss/ΔI; (c) Δm/ΔE. Source: Ref 20 The impedance diagram (Fig. 8a) shows two badly separated capacitive loops. The capacitance associated with the high-frequency loop is equal to 50 μF/cm2 and, therefore, can be attributed to the double-layer capacitance in parallel with the charge-transfer resistance. The low-frequency loop is then related to the faradaic impedance due to the relaxation of the surface coverage by a dissolution intermediate. In contrast with the case of iron presented above, Fig. 8(b) shows an emission efficiency N(f) = ΔIdiss/ΔI of 1 regardless of the perturbation frequency, indicating that there is no charge transfer in the formation of adsorbed intermediate species; that is, the reaction is purely chemical. The change in mass (Δm) induced by a perturbation in potential ΔE at frequency f gives a ratio Δm/ΔE parallel to the imaginary axis in the Nyquist plane (Fig. 8c) at low frequency, with a negative real part, which corresponds actually to electrode dissolution. By coupling EQCM and EIS data, the gram-equivalent (A), which represents the change in mass per mole of electrons exchanged (ΔQ/F), can be evaluated as follows:

(Eq 47)

since ΔI = jωΔQ and Z = ΔE/ΔI. The low- frequency limit of A(ω), -2πf0FIm(Δm/ΔE)Rp, can be obtained from the imaginary part of Δm/ ΔE = -33.5 × 10-6 μg/(V · cm2) at frequency f0 = 16 mHz. A value of 73.8 g

equivalent, close to the atomic mass of copper (63.5 g) when considering the experimental accuracy, was found, which confirms that copper actually dissolves as a monovalent ion in chloride medium at low anodic potential. The most probable dissolution mechanism consistent with the results of the EIS, RDE, and EQCM techniques is then the following: Cu + Cl- → Cu(Cl-)ad

(Eq 48)

Cu(Cl-)ad → (CuCl) + e-

(Eq 49)

The intermediate species, which is likely Cu(Cl-)ad, is adsorbed through a chemical reaction (no charge transfer), as suggested earlier. The subsequent electrochemical desorption of this species is potential dependent and generates the low-frequency capacitive loop in the impedance diagram. These combined EIS, RDE, and EQCM experiments allowed the electric state of the reaction intermediate to be demonstrated unambiguously. This is an important advancement in the understanding of the reaction mechanism.

Anodic Dissolution of a Binary Alloy: Iron-Chromium Iron-chromium (Fe-Cr) binary alloys constitute a model material for investigating corrosion of stainless steels. This is the reason why the dissolution and passivation mechanisms of iron- chromium alloys have been extensively studied in the literature. The mechanism of Fe-17Cr dissolution in 1M H2SO4 devised from EIS and RDE experiments (Ref 21) is discussed subsequently. Typical results are presented in Fig. 9. The polarization curve shows that this alloy becomes passive above -0.8 V versus SSE. At the critical potential of approximately -0.88 V versus SSE, the current density was about 3 mA/ cm2, several orders of magnitude lower than that measured with an iron electrode, indicating that chromium promotes the passivation process substantially.

Fig. 9 Dissolution of Fe-17Cr in 1 M H2SO4. (a) Polarization curve. (b) Impedance diagram at point A. (c) N(f) at point A. (d) ΔQ/ΔE at point A calculated from N(f) and Z data. Source: Ref 22

The impedance diagram, measured at point A of the polarization curve, exhibits four time constants, in addition to the high-frequency time constant related to the charge-transfer resistance in parallel with the double-layer capacitance. The four time constants correspond to the relaxation of the surface coverage by four intermediate species. Figure 9(c) presents the ratio N(f) = idiss/i where idiss is the current density relative to the dissolution of the Fe2+ ions collected at the ring and i is the overall current density. Both current densities are the responses to an ac potential perturbation at frequency f. It should be noted that i contains the contribution of both chromium and iron dissolution. A remarkable result is that the modulus of N(f) was greater than unity at frequencies above 1 Hz; that is, the faradaic efficiency of the iron dissolution was surprisingly greater than 100%, indicating that a surplus of charge was supplied to the iron dissolution, as explained later in this article. The ratio ΔQ/ΔE, where Q denotes the charge stored at the electrode surface calculated according to the algorithm developed for the dissolution of an iron electrode (Ref 17), is given in Fig. 9(d). To explain the main features of the aforementioned results and of those obtained in the passivation domain for alloys of various Cr content, the simplified reaction mechanism presented in Fig. 10 was devised. Only three reaction intermediates (Fe(I)ad, Fe(II)ad, and Cr(III)ad) were used to avoid a more intricate model that would be difficult to handle. The reactions relative to Cr contained the dissolution in Cr(II) sol species and the formation of a passive species Cr(III)ad. The Cr dissolution step is necessary to model the uniformity of the electrode dissolution, which gives a molar ratio of Fe and Cr dissolved in the solution equal to that of the metallic matrix. Otherwise, selective dissolution of iron would give rise to a macroporosity that was not observed experimentally. The effect of the Cr(III)ad passive species was to slow down the iron dissolution rate markedly, as proposed by Frankenthal (Ref 23). The mechanism of Fe dissolution contains the catalytic step (Eq B in Fig. 10) introduced by Heusler, but only two adsorbed species have been considered here instead of four in the mechanism proposed by Keddam and others (Ref 7). This model contains no Fe passivation reaction; the passive behavior observed experimentally was simulated by the Cr passivation that slows down the alloy dissolution by virtue of the uniform dissolution conditions.

Fe (Eq A) (Eq B) (Eq C) (Eq D) Cr (Eq E)

(Eq F)

Fig. 10 Schematic representation of the activation-passivation transition mechanism of Fe-Cr alloy in an acidic medium From the charge and mass balances for the reactions in Fig. 10, the polarization curve, the impedance diagrams, the dynamic emission coefficient N(f) of Fe2+, and the ratio ΔQ/ΔE were calculated. Figure 11 presents the results at point A of the polarization curve in Fig. 9(a). The impedance diagram was suitably reproduced except for the low-frequency behavior. This discrepancy is due to the simplifications in the model. The decrease in N(f) with decreasing frequency in Fig. 11(b) was also simulated, but the maximum value was significantly smaller than that obtained experimentally. The extra charge necessary to dissolve the iron beyond the Faraday law, mentioned earlier, was supplied by the surface relaxation of Cr(III)ad species. The ΔQ/ΔE curve was also correctly reproduced with two time constants and a negative real part, but the time constant in the simulated high-frequency loop was significantly smaller than the experimental one.

Fig. 11 Dissolution of Fe-17Cr in 1 M H2SO4. Simulations for the model presented in Fig. 10 at conditions of point A in Fig. 9. (a) Impedance diagram; (b) N(f); (c) ΔQ/ΔE. Source: Ref 21 The coupling of steady-state polarization, EIS, and RRDE experiments allowed a more reliable model of the reaction mechanism to be proposed. This is a very appropriate method to select the adequate model among all those devised from the single measurement of polarization curves.

References cited in this section 7. I. Epelboin and M. Keddam, J. Electrochem. Soc., Vol 117, 1970, p 1052 9. I. Epelboin, C. Gabrielli, M. Keddam, and H. Takenouti, Electrochemical Corrosion Testing, F. Mansfeld and U. Bertocci, Ed., STP 727, ASTM International, 1981, p 150–192 12. M. Keddam, Corrosion Mechanisms in Theory and Practice, P. Marcus, Ed., Marcel Dekker Inc, 2002, p 97 13. H. Takenouti, Electrochemistry, Vol 67, 1999, p 1063 14. K.E. Heusler, Z. Elektrochem., Vol 62, 1958, p 582 15. J.O'M. Bockris, D. Drazic, and A.R. Despic, Electrochim. Acta, Vol 4, 1961, p 325 16. M. Keddam, O. R. Mattos, and H. Takenouti, J. Electrochem. Soc., Vol 128, 1981, p 257, 266 17. N. Benzekri, M. Keddam, and H. Takenouti, Electrochim. Acta, Vol 34, 1989, p 1159 18. M. Itagaki, M. Tagaki, K. Watanabe, Denki Kagaku (Electrochemistry), Vol 63, 1995, p 425 19. O.E. Barcia, O.R. Mattos, and B. Tribollet, J. Electrochem. Soc. Vol 139, 1992, p 446 20. C. Gabrielli, M. Keddam, F. Minouflet-Laurent, and H. Perrot, Electrochem. Solid-State Letters, Vol 3, 2000, p 418 21. I. Annergren, M. Keddam, H. Takenouti, and D. Thierry, Electrochim. Acta, Vol 41, 1996, p 1121 22. R.P. Frankenthal, J. Electrochem. Soc., Vol 116, 1969, p 580 23. M. Keddam, O.R. Mattos, and H. Takenouti, Electrochim. Acta, Vol 31, 1993, p 1158


References 1. H. Kaesche, Metallic Corrosion, Principles of Physical Chemistry and Current Problems, NACE International, Houston, TX, 1985 2. F. Mansfeld and U. Bertocci, Ed., Electrochemical Corrosion Testing, STP 727, ASTM International, 1981

3. J.C. Scully, Ed., Treatise on Materials Science and Technology, Vol 23, Corrosion: Aqueous Processes and Passive Films, Academic Press, 1983 4. J.R. Scully, D.C. Silvermann, and M.W. Kendig, Ed., Electrochemical Impedance: Analysis and Interpretation, STP 1188, ASTM International, 1994 5. K.J. Vetter, Electrochemical Kinetics, Academic Press, 1967 6. A.J. Bard and L.R. Faulkner, Electrochemical Methods, Fundamentals and Applications, John Wiley & Sons, Inc., 1980 7. I. Epelboin and M. Keddam, J. Electrochem. Soc., Vol 117, 1970, p 1052 8. C. Wagner and W. Traud, Z. Elektrochem., Vol 44, 1938, p 391 9. I. Epelboin, C. Gabrielli, M. Keddam, and H. Takenouti, Electrochemical Corrosion Testing, F. Mansfeld and U. Bertocci, Ed., STP 727, ASTM International, 1981, p 150–192 10. M. Pourbaix, Lectures on Electrochemical Corrosion, Plenum Press, 1973 11. U.R. Evans, The Corrosion and Oxidation of Metals, Arnold Publications, Inc., 1960 12. M. Keddam, Corrosion Mechanisms in Theory and Practice, P. Marcus, Ed., Marcel Dekker Inc, 2002, p 97 13. H. Takenouti, Electrochemistry, Vol 67, 1999, p 1063 14. K.E. Heusler, Z. Elektrochem., Vol 62, 1958, p 582 15. J.O'M. Bockris, D. Drazic, and A.R. Despic, Electrochim. Acta, Vol 4, 1961, p 325 16. M. Keddam, O. R. Mattos, and H. Takenouti, J. Electrochem. Soc., Vol 128, 1981, p 257, 266 17. N. Benzekri, M. Keddam, and H. Takenouti, Electrochim. Acta, Vol 34, 1989, p 1159 18. M. Itagaki, M. Tagaki, K. Watanabe, Denki Kagaku (Electrochemistry), Vol 63, 1995, p 425 19. O.E. Barcia, O.R. Mattos, and B. Tribollet, J. Electrochem. Soc. Vol 139, 1992, p 446 20. C. Gabrielli, M. Keddam, F. Minouflet-Laurent, and H. Perrot, Electrochem. Solid-State Letters, Vol 3, 2000, p 418 21. I. Annergren, M. Keddam, H. Takenouti, and D. Thierry, Electrochim. Acta, Vol 41, 1996, p 1121 22. R.P. Frankenthal, J. Electrochem. Soc., Vol 116, 1969, p 580 23. M. Keddam, O.R. Mattos, and H. Takenouti, Electrochim. Acta, Vol 31, 1993, p 1158

F. Huet, R.P. Nogueira, and H. Takenouti, Aqueous Corrosion Reaction Mechanisms, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 52–60 Aqueous Corrosion Reaction Mechanisms

François Huet, Ricardo P. Nogueira, Bernard Normand, and Hisasi Takenouti, Université Pierre et Marie Curie and UPR 15 of CNRS, “Laboratoire Interfaces et Systèmes Electrochimiques,” Université Pierre et Marie Curie

Selected References • • • • •

A.J. Bard and L.R. Faulkner, Electrochemical Methods, Fundamentals and Applications, John Wiley & Sons, Inc., 1980 C. Gabrielli, “Identification of Electrochemical Processes by Frequency Response Analysis,” Solartron Instruments, Farnborough, U.K., 1980 C. Gabrielli, “Use and Applications of Electrochemical Impedance Techniques”, Technical Report 24, Solartron Instruments, Farnborough, U.K., 1997 H. Kaesche, Metallic Corrosion, Principles of Physical Chemistry and Current Problems, NACE International, 1985 M. Keddam, Anodic Dissolution, Corrosion Mechanisms in Theory and Practice, P. Marcus, Ed., Marcel Dekker Inc., 2002, p 97

J. Kruger, Passivity, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 61–67

Passivity Jerome Kruger, Johns Hopkins University

Introduction ALL METALS AND ALLOYS, with the exception of gold, have a thin protective corrosion product film present on their surface resulting from reaction with the environment. If such a film did not exist on metallic materials exposed to the environment, they would revert back to the thermodynamically stable condition of their origin—the ores used to produce them. Some of these films—the passive films—on some, but not all, metals and alloys have special characteristics that enable them to provide superior corrosion- resistant metal surfaces. These protective “passive” films are responsible for the phenomenon of passivity. The first metal found to exhibit the phenomenon of passivity was iron. Uhlig (Ref 1) has written a review of the history of passivity that lists three 18th century scientists—the Russian Lomonosov in 1738, the German Wenzel in 1782, and the Briton Keir in 1790—who observed that the highly reactive surface of iron became unexpectedly unreactive after immersion in concentrated nitric acid. This effect was first called passivity by Schönbein. This unexpected phenomenon of passivity occupies a central position in controlling corrosion processes, enabling the use of metallic materials in the many technologies of the 21st century. Moreover, it is the breakdown of the passive film that leads to the inability of metals and alloys to perform their assigned functions because of localized corrosion failure modes such as stress corrosion, pitting, crevice corrosion, and corrosion fatigue. Its importance to materials technology transcends, however, corrosion science and corrosion engineering. For example, one of the main reasons silicon replaced germanium in semiconductor device technology was that silicon forms effective passive films and germanium does not (Ref 2). Early work in the area of passivity that had an enormous impact on providing technology with improved engineering materials is, of course, the development of the stainless steels. This has promoted the continual development of a large number of alloys that exhibit corrosion resistance because of the protection provided by the passive film.

An improved understanding of the role that alloying constituents play in determining the properties of this passive film will lead to guidelines that can be used to develop engineering alloys with improved corrosion resistance. The scope of this article limits the discussion of all of the details on the subject of passivity. Moreover, the passivation behavior of all of the various metals and semiconductors that exhibit passivity is not given. Instead, this article discusses the classic passive metal iron and its alloys as illustrative examples of metals exhibiting passivity. References 3, 4, 5, 6, 7 provide a more extensive treatment of the subject of passivity in general and passivity of other metals and semiconductors in addition to iron.

References cited in this section 1. H.H. Uhlig, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 1–28 2. A.G. Revesz and J. Kruger, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 137 3. R.P. Frankenthal and J. Kruger, Ed., Passivity of Metals, Electrochemical Society, 1978 4. J. Kruger, Int. Mater. Rev., Vol 33, 1988, p 113–130 5. H. Hasegawa and T. Sugano, Ed., Passivation of Metals and Semiconductors, Part II, Passivity of Semiconductors, Pergamon Press, 1990 6. K.E. Heusler, Ed., Passivation of Metals and Semiconductors, Materials Science Forum, Vol 185–188, Trans Tech Publications, 1995 7. M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Passivity of Metals and Semiconductors, Proc. Vol 99–42, Electrochemical Society, 2001

J. Kruger, Passivity, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 61–67 Passivity Jerome Kruger, Johns Hopkins University

General Aspects Importance of Passivity to Corrosion-Control Technology. If the passive film did not exist, most of the technologies that depend on the use of metals could not exist because the phenomenon of passivity is a critical element in controlling corrosion processes. Therefore, the destruction of passivity at local breakdowns leads to a large part of the corrosion failures of metal and alloy structures—localized attack such as pitting, crevice corrosion, stress corrosion, and corrosion fatigue. The development of the stainless steels in the 1920s is regarded as a major application of the phenomenon of passivity. This development has contributed significantly to modern technology by providing the design engineer with engineering materials such as the large number of iron and nickel-base alloys as well as many other alloy systems that exhibit superior corrosion resistance—this effort continues today. Types of Passivity. There are two types of passivity:

•

•

Type 1. A metal active in the electromotive force (emf) series is passive when its electrochemical behavior in a given environment becomes that of a metal noble in the emf series (low corrosion rate, noble potential). Type 2. A metal is passive while, still from the standpoint of thermodynamics at an active potential in a given environment, it exhibits a low corrosion rate (low corrosion rate, active potential). This type of passivity can be termed “practical passivity.”

Only type 1 passivity is considered here. Examples of metals or alloys exhibiting such passivity are nickel, chromium, titanium, iron in oxidizing environments, stainless steels, and many others. Examples of type 2 passivity are lead in sulfuric acid and iron in an inhibited pickling acid. A major characteristic of a type 1 passive system is the existence of a polarization curve (i, current density, or rate, versus E, potential, or driving force), of the sort shown in Fig. 1. It illustrates well a restatement of the definition of type 1 passivity as first proposed by Wagner (Ref 8). He suggested that a metal becomes passive when, upon increasing its potential in the positive or anodic (oxidizing) direction, a potential is reached where the current (rate of anodic dissolution) sharply decreases to a value less than that observed at a less anodic potential. This decrease in anodic dissolution rate, in spite of the fact that the driving force for dissolution is brought to a higher value, is the result of the formation of a passive film.

Fig. 1 The idealized anodic polarization curve for an iron-water system exhibiting passivity. Three different potential regions are shown; the active, passive, and pitting or transpassive regions. Ep is potential above which the system becomes passive and exhibits the passive current density ip. The critical current density for passivation is ic. Another more practical definition has been provided by an ASTM standard: “passive—the state of metal surface characterized by low corrosion rates in a potential region that is strongly oxidizing for the metal” (Ref 9). Employing Passivity to Control Corrosion. Passivity can be used to control corrosion by using methods that bring the potential of the surface to be protected to a value in the passive region. This can be accomplished by the following tactics: •

Using a device called a potentiostat, a current can be applied to the metal to be protected that will set and control the potential at a value greater than the passivating potential, Ep. This method of producing passivity is called anodic protection (Fig. 1).

•

•

•

•

For environments containing damaging species such as chloride ions that cause pitting, the potentiostat or other devices that control the potential can be used as in the item above to set the potential to a value in the passive region below the critical potential for pitting, Epit. Alloys or metals that spontaneously form a passive film, for example, stainless steels, nickel, or titanium alloys, can be used in applications that require resistance to corrosion. Usually a pretreatment such as that described below is desirable. A surface pretreatment can be carried out on an alloy capable of being passivated. The use of such a pretreatment has been a standard practice for stainless steels for many years. The passivating procedure involves immersion of thoroughly degreased stainless steel parts in a nitric acid solution followed by a thorough rinsing in clean, hot water. The most popular solution and conditions of operation for passivating stainless steel is a 30 min immersion in a 20 vol% nitric acid solution at 49 °C (120 °F). However, other solutions and treatments may be used, depending on the type of stainless steel being treated (Ref 10). The environment can be modified to produce a passive surface. Oxidizing agents such as chromate and concentrated nitric acid are examples of passivating solutions that maintain a passive state on some metals and alloys.

References cited in this section 8. C. Wagner, Corros. Sci., Vol 5, 1965, p 751–764 9. Definitions of Terms Relating to Corrosion and Corrosion Testing, G 13, Annual Book of ASTM Standards, ASTM, 1983 10. D. Peckner and I.M. Bernstein, Handbook of Stainless Steels, McGraw-Hill, 1977, Ch 24, p 24


Thermodynamics of Passivity Thermodynamics provide a guide to the conditions under which passivation becomes possible. A valuable guide to thermodynamics is the potential-pH diagram, the Pourbaix diagram. Pourbaix's Atlas of Electrochemical Equilibria in Aqueous Solutions (Ref 11) describes applications of these potential-pH equilibrium diagrams to corrosion science and engineering. One major application is the establishment of the theoretical domains or conditions for corrosion, immunity, and passivation. Figure 2 shows a simplified diagram for the iron-water system. The three theoretical domains show on a thermodynamic basis the potential- pH conditions where no corrosion is possible (immunity), where a corrosion-product film forms that may confer protection against corrosion (passivation), and where corrosion is expected (corrosion). (Pourbaix designates the immunity domain as that of “thermodynamic nobility” and the total of the passivation and immunity domains as that of “practical nobility.”) Whether the film is passive (protective) or not is a kinetic consideration and not a thermodynamic one (see the section “Nature of the Passive Film” in this article). Such Pourbaix diagrams can identify metals capable of forming films that, depending on their properties, may or may not be protective, and conditions can be determined where there is a transition from passivation to activation. One could call the equilibrium diagrams a “road map of the possible.”

Fig. 2 Simplified potential-pH equilibrium diagram (Pourbaix diagram) for the iron-water system. Above equilibrium line A oxygen is evolved, and below equilibrium line B hydrogen is evolved. Source: Ref 11 The diagrams can, therefore, be used as a basis for identifying the active, passive, and transpassive regions of active-passive polarization curves (see Fig. 1). Thus, potentials above the oxygen-evolution line (the line marked A in Fig. 2) are in the transpassive region. Also, the diagrams can be used to interpret the reasons for loss of the protective nature of the passive film in the transpassive region. For example, the protective layer on stainless steels that contain chromium involves Cr(III); at higher potentials Cr(III) is oxidized to Cr(VI), and the protective Cr2O3 becomes the soluble chromate ion, resulting in the loss of corrosion resistance. Usually, the passive regions of the polarization curves correspond to potentials in the equilibrium diagrams where protective solid compounds are stable. However, even though the active regions of the polarization curves usually lie in regions labeled as “corrosion” on the potential-pH diagrams, this is not always the case. For example, iron can passivate in sulfuric acid solutions under conditions where the diagrams would predict corrosion and, hence, an active condition, but the rate of passive-film dissolution is so extremely slow that the film is metastable and thereby prevents metal dissolution.

Reference cited in this section 11. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, 2nd ed., Pergamon Press, 1966


Kinetics of Passivity From the standpoint of the kinetics, passivity can be characterized as the conditions existing on a metal surface because of the presence of a protective film that markedly lowers the rate of corrosion, even though from thermodynamic (corrosion tendency) considerations one would expect active corrosion. Figure 1, which depicts

an idealized anodic polarization curve for a metal surface, can serve as a basis for describing in a general way the kinetics of passivation. Anodic polarization curves obtained from a real system under practical conditions (Ref 12) are shown in Fig. 3. They still include the general features of the ideal curve (Fig. 3). Figure 3 shows a comparison of iron and 304L stainless steel in H2SO4. In Fig. 1, where the anodic polarization curve is that of a metal that exhibits an ability to become passive, the current initially increases with an increase in potential, but when the potential reaches the value of the passivating potential, Ep, the critical current density for passivation, ic, is reached, and a marked drop in current density (corrosion rate) is observed. This is the onset of passivity, and the current density remains low at ip as the potential is increased to higher values. If the potential is increased to sufficiently high values, the current density begins to rise, and either pitting results or the transpassive region is entered. In the transpassive region, oxygen evolution and possibly increased corrosion takes place.

Fig. 3 Comparison of anodic polarization curves for iron and 304L stainless steel in 1 N H2SO4. Adapted from Ref 12 The corrosion potential of a metal surface is controlled by the intersection of the anodic (potential increases in the positive direction) and cathodic (potential increases in the negative direction) polarization curves where the anodic and cathodic reaction rates are equal. Therefore, even though a metal may be capable of exhibiting passivity, its corrosion rate will depend on where the cathodic polarization curve intersects the passive metal anodic curve of the type shown in Fig. 1. Figure 4 shows three possible cases. If the cathodic reaction produces a polarization curve such as A, which is indicative of oxidizing conditions, the corrosion potential will be located in the passive region, and the system can exhibit a low corrosion rate. If the cathodic reaction produces curve C, which is indicative of reducing conditions, the corrosion potential will be in the active region, and the corrosion rate can be high. Curve B represents an intermediate case where passivity, if it exists at all, will be unstable, and the surface will oscillate between active and passive states.

Fig. 4 Intersections of three possible cathodic polarization curves (straight lines A, B, C) with an anodic polarization curve for a system capable of exhibiting passivity. The corrosion rate depends on the current density at the intersection. Curve A produces a passive system, curve C an active system, and curve B an unstable system.

Reference cited in this section 12. M.G. Fontana and N.D. Greene, Corrosion Engineering, McGraw-Hill, 1967, p 336–337


Nature of the Passive Film It is now widely believed that a film is responsible for the condition of passivity—one of the major accomplishments of past research. The understanding of the nature of passive films has been greatly enhanced in recent years, resulting in the development of many models of the passive film by the application of a whole array of in situ and ex situ techniques developed over the past 25 to 30 years. Two examples of collections of such models (Fig. 5) have been given by Sato (Ref 13) and Cohen (Ref 14). This section discusses five of the properties of passive films: the thickness, the composition, the structure, and the electronic and mechanical properties. These five aspects are discussed using, for the purpose of illustration, the films on iron and iron alloys. Since all of the properties that determine the nature of the passive film are interrelated, the discussion of each is artificially limited.

Fig. 5 Proposed models of the passive film. (a) General models include monolayers and multiple layers. Source: Ref 13. (b) Detailed proposed models for iron having single or double layers containing combinations of oxides, hydroxides, and oxyhydroxides. Source: Ref 14 Thickness. Ever since passivity was discovered, the thickness of the passive film has been a source of great controversy between those who proposed a two-dimensional film (Ref 15), in which an adsorbed monolayer or less than a monolayer of oxygen retards surface reaction rates, and those who proposed a three-dimensional film (Ref 16), in which a phase oxide that had a thickness greater than one unit cell could serve as a barrier to the diffusion of metal cations into the solution. The application of numerous techniques appear to have resolved this issue because, depending on conditions, it has been shown by a number of studies that the passive film can be either two- or three-dimensional. Some of these studies establishing the three- dimensional picture (thickness > monolayer) have been: • •

In situ ellipsometric (Ref 17, 18) measurements of the thickness of the passive film as it formed Cathodic reduction (coulometric) measurements of thickness (Ref 19, 20)

For confirming passive-film thicknesses of less than a monolayer, coulometric electrochemical techniques have been employed (Ref 21, 22, 23). One of these studies by Frankenthal (Ref 22) have actually provided a link between two- and three-dimensional films. He found that at low potentials in the passive region (-0.4 to -0.1 V(SHE)) for iron in a nearly neutral borate-buffer solution, the film measured was less than a unit cell for Fe3O4 or γFe2O3 (around 0.84 nm). This film may be considered to be adsorbed oxygen. Above these potentials, he measured thicknesses greater than the unit cell for the phase oxide with Fe(III), for example, γFe2O3. Ellipsometric studies (Ref 24) on easily passivated metals such as chromium also show a quite wide potential region (as much as 1000 mV wide) where passive films exhibit thicknesses less than one unit cell of a phase oxide both in neutral and acidic solutions. Chemical composition is, perhaps, the major property controlling the nature of passive films. For the passive film on iron, many studies have resulted in a confusing array of chemical compositions. A good summary of some of the ideas that have come out of these investigations is the collection of some of the proposed models shown in Fig. 5(b) (Ref 14). These models involve either single or double layers that contain different combinations and arrangements of the following oxides, hydroxides, or oxyhydroxides: Fe3O4, γFe2O3, FeOOH, a polymeric layered Fe(OH)2 (Ref 25), a nonstoichiometric cation-deficient γFe2O3 containing varying amounts of protons (Fe2-xHxO3) (Ref 26), and a cation-deficient Fe2O3 (Fe2-2xGxO3) (Ref 14). In addition to this list from the two models shown in Fig. 5(b) is a later model from Cahan and Chen (Ref 27), who characterize the chemical composition loosely as “a highly protonated, trivalent iron oxyhydroxide capable of existing over a relatively wide range of stoichiometry.” The models for iron shown in Fig. 5(b) have led to a number of issues concerning the chemical compositions of passive films: • • •

The number of layers in a passive film (Ref 20, 25, 26, 28, 29, 30, 31, 32, 33) The presence of hydrogen in some passive films (Ref 30, 34, 35, 36, 37, 38, 39, 40) (where in situ techniques are necessary) The existence, nature, and binding states of alloying elements with oxygen in the passive films on alloys (Ref 31, 33, 41, 42, 43)

Structure. Because chemical composition determines structure, these two aspects of the nature of the passive film are tied closely together, making much of the discussion on chemical composition relevant to structure. A major emphasis of many of the structural investigations of the structure of the passive film has been on the issue of crystallinity. Depending on the metal or alloy bearing the passive layer, ex situ studies (Ref 29, 32, 44, 45)—some researchers have found (Ref 32, 34) that structural changes may take place upon the transfer of a passivated specimen from an aqueous solution to the vacuum used in an ex situ technique—and in situ studies (Ref 30, 34, 39, 45, 46) found passive films with either crystalline or noncrystalline structures. Some studies (Ref 47, 48, 49, 50) have contended that for some systems, for example, high-chromium stainless steels or passive films formed on iron by passivation in chromate solutions, the film is noncrystalline and that this noncrystallinity is promoted by certain alloying elements such as chromium and by the presence of

hydrogen in the structure of the film (Ref 2). There is also a large body of literature suggesting that the crystallization of the oxide layers on titanium alloys adversely affects the properties that enhance the passivity of the film (Ref 51, 52, 53, 54). Other studies (Ref 55, 56) have, however, found that superior passive films are crystalline and become more so upon aging (Ref 57). Electronic Properties. This aspect of the nature of the passive film is an important factor in controlling the mechanisms of film formation, breakdown, and the rate of metal dissolution. This is so because dissolution, film formation, and breakdown all involve the movement of electrons and ions from the metal surface through the passive film or from the solution into the film. Moreover, electron-transfer reactions that occur on surfaces with passive films depend strongly on the electronic properties of such films. Iron—like other metals exhibiting type 1 passivity but unlike electronic valve metals such as aluminum, titanium, and tantalum—forms a very thin, passive layer (less than 10 nm). The valve metals, however, whose films are good insulators, can support large electric fields and by so doing form quite thick films (hundreds of nanometers). Oxygen cannot be evolved from valve metals. Iron, when high potentials are applied, evolves oxygen instead of continuing to grow a thicker film. It is for this reason that many workers have suggested that the passive film on iron is a good electronic conductor, or at least a semiconductor (Ref 58). Many ideas on the role of electronic properties of electron-transfer reactions at the passive-film surface have suggested films with different electronic characteristics, namely a semiconducting film (Ref 58, 59, 60, 61, 62, 63) or a film with low electronic conductivity that is an insulator or partially an insulator to support the large fields required if the proposed mechanism of film growth is field-assisted ionic migration (Ref 19, 50, 64, 65). Cahan and Chen (Ref 27) attempted to reconcile these opposing findings, proposing that the passive film on iron is neither a semiconductor nor an insulator, but a combination of both, that is, a “chemi-conductor,” which they define as “a material whose stoichiometry can be varied by oxidative and/or reductive valency state changes. This nonstoichiometry can then modify the local electronic (and/or ionic) conductivity of the film.” The mechanical properties of passive films can be an important factor in the breakdown of passivity. Even though the determination of the mechanical properties of passive films is difficult, a few attempts have been made to measure these properties for a number of metals (Ref 66, 67, 68). These studies have shown that applied potentials (Ref 68) and alloying (Ref 66, 67) can control the ductility of some passive layers. It has been suggested (Ref 2) that the effect of alloying may be a consequence of the passive film, for example, on a chromium alloy being more noncrystalline than that on iron. An ex situ study of the passive layers formed in nitric acid solutions on stainless steels (Ref 52) found these films to be crystalline, epitaxial, and composite. This suggested that the mechanical properties of such films may be anisotropic and brittle with a high degree of adherence stress at the metal/film interface.

References cited in this section 2. A.G. Revesz and J. Kruger, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 137 13. N. Sato, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 29– 58 14. M. Cohen, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 521–545 15. H.H. Uhlig, Z. Elektrochem., Vol 62, 1958, p 626–632 16. U.R. Evans, J. Chem. Soc., Vol 1927, 1927, p 1020–1040 17. J. Kruger, J. Electrochem. Soc., Vol 110, 1963, p 654–663 18. J. Kruger and J.P. Calvert, J. Electrochem. Soc., Vol 114, 1967, p 43–49 19. P.M.G. Draper, Corros. Sci., Vol 7, 1967, p 91–101

20. M. Nagayama and M. Cohen, J. Electrochem. Soc., Vol 109, 1962, p 781–790 21. K. Kubanov, R. Burstein, and A. Frumkin, Disc. Faraday Soc., Vol 1, 1947, p 259–269 22. R.P. Frankenthal, Electrochim. Acta, Vol 16, 1971, p 1845–1857 23. R.P. Frankenthal, J. Electrochem. Soc., Vol 114, 1967, p 542–547 24. K.E. Heusler, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 771–801 25. R.W. Revie, B.G. Baker, and J.O'M. Bockris, J. Electrochem. Soc., Vol 122, 1975, p 1460–1466 26. M.C Bloom and M. Goldenberg, Corros. Sci., Vol 5, 1965, p 623–630 27. B.D. Cahan and C.-T. Chen, J. Electrochem. Soc., Vol 129, 1982, p 921–925 28. K.S. Vetter, Z. Elektrochem., Vol 62, 1958, p 642–648 29. C.L. Foley, J. Kruger, and C.J. Bechtoldt, J. Electrochem. Soc., Vol 114, 1967, p 944–1001 30. H.T. Yolken, J. Kruger, and J.P. Calvert, Corros. Sci., Vol 8, 1968, p 103–108 31. K. Asami, K. Hashimoto, T. Musumoto, and S. Shimodaira, Corros. Sci., Vol 16, 1976, p 909–914 32. K. Kuroda, B.D. Cahan, Ch. Nazri, E. Yeager, and T.E. Mitchell, J. Electrochem. Soc., Vol 129, 1982, p 2163–2169 33. L.J. Jablonsky, M.P. Ryan, and H.S. Isaacs, J. Electrochem. Soc., Vol 145, 1998, p 1922–1932 34. W.E. O'Grady, J. Electrochem. Soc., Vol 127, 1980, p 555–563 35. G. Okamoto and T. Shibata, Nature, Vol 206, 1965, p 1350 36. G.G. Long, J. Kruger, D.R. Black, and M. Kuriyama, J. Electrochem. Soc., Vol 130, 1983, p 240–242 37. M. Kekar, J. Robinson, and A.J. Forty, Faraday Discuss. Chem. Soc., Vol 89, 1990, p 31–40 38. A.J. Davenport, and M. Sansone, J. Electrochem. Soc., Vol 142, 1995, p 727–730 39. J. Kruger, L.A. Krebs, G.G. Long, J.F. Anker, and C.F. Majkrzak, Passivation of Metals and Semiconductors, K.E. Heusler, Ed., Materials Science Forum, Vol 185–188, Trans Tech Publications, 1995, p 367–376 40. C.R. Clayton, K. Doss, and J.B. Warren, Passivity of Metals and Semiconductors, M. Froment, Ed., 1983, p 585–590 41. A.J. Davenport, H.S. Isaacs, J.A. Bardwell, B. MacDougall, G.S. Frankel, and A.G. Schrott, Corros. Sci., Vol 35, 1993, p 19–25 42. A.J. Davenport, M. Sansone, J.A. Bardwell, A.J. Andlykiewicz, M. Taube, and C.M. Vitus J. Electrochem. Soc., Vol 141, 1994, p L6–8 43. J. Eldrige, M.E. Kordesch, and R.W. Hoffman, J. Vac. Sci. Technol., Vol 20, 1982, p 934–938

44. G.G. Long, J. Kruger, D.R. Black, and M. Kuriyama, J. Electroanal. Chem., Vol 150, 1983, p 603–610 45. L.J. Oblonsky, A.J. Davenport, M.P. Ryan, and M.F. Toney, Passivity of Metals and Semiconductors, M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Proc. Vol 99-42, Electrochemical Society, 2001, p 173–177 46. A.J. Davenport, R.C. Newman, and P. Ernst, Passivity of Metals and Semiconductors, M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Proc. Vol 99-42, Electrochemical Society, 2001, p 65–71 47. T.P. Hoar, J. Electrochem. Soc., Vol 117, 1970, p 17C–22C 48. M.P. Ryan, R.C. Newman, and G.E. Thompson, Philos. Mag. B, Vol 70, 1994, p 241–251 49. M.P. Ryan, S. Fugimoto, G.E. Thompson, and R.C. Newman, Passivation of Metals and Semiconductors, K.E. Heusler, Ed., Materials Science Forum, Vol 185–188, Trans Tech Publications, 1995, p 233–240 50. G.G. Long, J. Kruger, M. Kuriyama, D.R. Black, E. Farabaugh, D.M. Saunders, and A.I. Goldman, Proc. Ninth Int. Congress on Metallic Corrosion, Proc. Vol 3, National Research Council, Ottawa, Canada, 1984, p 419–422 51. T. Shibata and Y.-C. Zhu, Corros. Sci., Vol 36, 1994, p 153–163 52. M. Pankuch, R. Bell, and C.A. Melendres, Electrochim. Acta, Vol 38, 1993, p 2777–2779 53. K.E. Healy and P. Ducheyne, Mater. Sci., Vol 4, 1993, p117–126 54. J.S.L. Leach and B.R. Pearson, Corros. Sci., Vol 28, 1988, p 43–56 55. P. Marcus and V. Maurice, Passivity of Metals and Semiconductors, M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Proc. Vol 99-42, Electrochemical Society, 2001, p 30–54 56. V. Vignal, J.M. Olive, and J.C. Roux, Passivity of Metals and Semiconductors, M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Proc. Vol 99-42, Electrochemical Society, 2001, p 208–213 57. W. Yang, D. Costa, and P. Marcus, J. Electrochem. Soc., Vol 141, 1994, p 2669–2676 58. K.J. Vetter, J. Electrochem. Soc., Vol 110, 1963, p 597–605 59. F.M. Delnick and N. Hackerman, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 116–133 60. E.K. Oshe, I.L. Rosenfeld, and V.C. Doizoskenko Dokl. Akad. Nauk SSSR, Vol 194, 1970, p 612–614 61. M. Bojinov, T. Laitinen, K. Mäkelä, T. Saario, T. Sirkiä, and G. Fabricius, Passivity of Metals and Semiconductors, M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Proc. Vol 99-42, Electrochemical Society, 2001, p 201–207 62. W. Schmickler, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 102–115 63. W. Schultze and U. Stimming, Z. Phys. Chem., NF, Vol 98, 1975, p 285–302 64. J. Ord and D.J. DeSmet, J. Electrochem. Soc., Vol 113, 1966, p 1258–1262

65. R.V. Moshtev, Electrochim. Acta., Vol 16, 1971, p 2039–2048 66. S.F. Bubar and D.A. Vermilyea, J. Electrochem. Soc., Vol 113, 1966, p 892–895 67. S.F. Bubar and D.A. Vermilyea, J. Electrochem. Soc., Vol 114, 1967, p 882–885 68. J.S.L. Leach and P. Neufeld, Corros. Sci., Vol 9, 1969, p 225–244


Passive-Film Formation and Dissolution Processes. Passivity results from bringing a metal surface into the passivity region of a system exhibiting a passive polarization curve (Fig. 1) and thereby forming “the passive film.” Sato and Okamoto (Ref 69) have pointed out that there are three possible processes that can produce passive films. The three passivation processes defined below describe possible ways in which the imposition of an electrochemical current can result in the formation of a passive film. Direct film formation involves the reaction of a metal surface with an aqueous solution to form either a chemisorbed oxygen film or a compact three-dimensional (more than one atomic layer) film, usually an oxide or oxyhydroxide represented by: M + H2O = MOads + 2H+ + 2e-

(Eq 1)

or M + H2O = MO(oxide) + 2H+ + 2e-

(Eq 2)

Dissolution precipitation produces a passive layer by the formation of an oxide, oxyhydroxide, or hydroxide film by the precipitation of dissolved metal ions as described by the two-step process: M = Mz+ + ze-

(Eq 3)

Mz+ + zH2O = M(OH)z + zH+

(Eq 4)

The anodic oxidation of metal ions in solution forms an oxide film containing the metal ion in a higher oxidation state as shown by: M = Mz+ + ze-

(Eq 5)

2Mz+ + (z + x)H2O = M2O(z+x) + 2(z + x)H+ + 2xe-

(Eq 6)

Rate Laws of Passive-Film Formation. As a result of the passivation processes described previously, it is possible to develop various expressions for the rate of passive-film formation by applying either of two electrochemical methods to form the passive film; these methods are described below. Galvanostatic anodic oxidation or passivation applies a constant current and measures the change in potential as a function of time. When the potential for that surface, with respect to a reference electrode, is observed as a function of time, it is found that the potential rises initially during an induction time, τp, until it reaches a relatively constant value, Ep, the passivation potential, as shown in Fig. 1. The relationship between τp and the critical current density for passivation, ic, is given by Sato and Okamoto (Ref 69) as τp = k(i - ic)n, where k and n are constants with n having been found to be -1 for iron by Frank (Ref 70).

Potentiostatic anodic oxidation or passivation sets the potential of the metallic surface at a constant value and observes the variation of the current with time. It has been proposed (Ref 71) that potentiostatic passivation involves a competition between metal dissolution and film formation with the total current density, which is the reaction rate for the passivation process given by the expression i = (idiss + ifilm) (1 - θ), where idiss is the current density for film dissolution, ifilm is the current density for film formation, and θ is the fraction of the surface covered by the passive film. Mechanism of Formation. A review of the mechanisms of passive-film formation has been given by Fromhold (Ref 72). He has pointed out that it has been shown experimentally that the relationship between the electric field E across the passive film (potential difference across the film divided by the film thickness, usually several millions volts per centimeter) and the current density, i, can be given by: E - Eo = A log i

(Eq 7)

where Eo and A are constants. A number of the proposed mechanisms for the growth of the passive film in the limiting thickness region (where film growth levels off) are some form of a field- assisted ion conduction mechanism based on the oxidation theory developed by Cabrera and Mott (Ref 73) such as the hopping (Ref 74), induced space charge (Ref 75), and point defect (Ref 76) mechanisms. A major problem in deciding which of the field-assisted ion conduction mechanisms is operative for iron in neutral solutions is that the kinetics of passive-film growth follow equally well either inverse logarithmic or direct logarithmic rate laws. Figure 6 shows the extent of this problem for iron in a neutral borate buffer solution. The inverse logarithmic law indicates a field-assisted ion conduction mechanism; the direct logarithmic law can be expected from a placeexchange mechanism. In addition, there are the models that do not depend on field-assisted ion conduction. These include the chemisorption of oxygen model (Ref 15, 77), the place-exchange mechanism (Ref 78), and the bipolar fixed charge induced passivity mechanism (Ref 79). Fromhold (Ref 72) has found that no mechanism can adequately explain all aspects of the film-formation process.

Fig. 6 Logarithmic plots of the growth of passive film on iron by potentiostatic anodic polarization at different potentials in pH 8.4 borate-buffer solution (a) Direct. (b) Inverse. Source: Ref 70 Dissolution. The process of passive-film dissolution is as important as that of film formation in controlling corrosion. Attention has mainly been placed on the dissolution of passive films in acid solutions (Ref 80) that involve either galvanostatic or steady-state open circuit conditions (Ref 58). An important aspect of passivefilm dissolution is the existence of a potential, the Flade potential, that delineates the transition from the passive to the active state. It was first observed by Flade (Ref 81) when he found that the open-circuit potential of a passive metal surface decreased continuously and then ceased to change momentarily and arrested, before decaying to more active values that signaled the onset of the active state. Uhlig and King (Ref 82) have given a number of examples of this transition through the Flade potential during the decay of passivity. Other studies (Ref 33, 42, 83, 84) focused on the examination of the reduction and dissolution of individual species in passive films in mildly acidic and basic solutions. Finally, an investigation (Ref 85) that studied passive-film

dissolution in nearly neutral solutions was able to distinguish between the field and chemical effects that result in passive-film thinning.

References cited in this section 15. H.H. Uhlig, Z. Elektrochem., Vol 62, 1958, p 626–632 33. L.J. Jablonsky, M.P. Ryan, and H.S. Isaacs, J. Electrochem. Soc., Vol 145, 1998, p 1922–1932 42. A.J. Davenport, M. Sansone, J.A. Bardwell, A.J. Andlykiewicz, M. Taube, and C.M. Vitus J. Electrochem. Soc., Vol 141, 1994, p L6–8 58. K.J. Vetter, J. Electrochem. Soc., Vol 110, 1963, p 597–605 69. N. Sato and G. Okamoto, Comprehensive Treatise of Electrochemistry, J.O'M. Bockris et al., Ed., Vol 4, Plenum Press, 1981, p 193–306 70. U.F. Frank, Z. Naturforsch., Vol 4A, 1949, p 378–391 71. U. Ebersbach, K. Schwabe, and K. Ritter, Electrochim. Acta, Vol 12, 1967, p 927–938 72. A.T. Fromhold, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 59–81 73. N. Cabrera and N. Mott, Rep. Prog. Phys., Vol 12, 1949, p 163–184 74. J. Kruger and J.P. Calvert, J. Electrochem. Soc., Vol 114, 1967, p 43–49 75. A.T. Fromhold, Jr., and J. Kruger, J. Electrochem. Soc., Vol 120, 1973, p 722–729 76. C.Y. Chao, L.F. Lin, and D.D. Macdonald, J. Electrochem. Soc., Vol 128, 1981, p 1187–1194 77. Ya.M. Kolotyrkin and V.M. Knyazheva, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 678–698 78. N. Sato and M. Cohen, J. Electrochem. Soc., Vol 111, 1964, p 513–519 79. M. Sakashita and N. Sato, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 479–483 80. K.J. Vetter and F. Gorn, Electrochim. Acta, Vol 18, 1973, p 321–326 81. F. Flade, Z. Phys. Chem., Vol 76, 1911, p 513–559 82. H.H. Uhlig and P.F. King, J. Electrochem. Soc., Vol 106, 1959, p 1–7 83. A.J. Davenport, H.S. Isaacs, J.A. Bardwell, B. MacDougall, G.S. Frankel, and A.J. Schrott, Corros. Sci., Vol 35, 1993, p 19–25 84. L.J. Oblonsky, Passivity of Metals and Semiconductors, M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Proc. Vol 99-42, Electrochemical Society, 2001, p 253–257 85. J. Kruger, Proc. Corrosion and Corrosion Protection, R.P. Frankenthal and F. Mansfeld, Ed., Vol 81-8, Electrochemical Society, 1981, p 66–76


Breakdown of the Passive Film When a fourth cathodic curve is added above line A in Fig. 4, a line that intersects the anodic passive curve at a point where the current density is increasing sharply, breakdown of the passive film results, leading to the most damaging kinds of corrosion—the localized forms of corrosion, pitting, crevice corrosion, intergranular attack, and stress corrosion. A number of reviews of breakdown or pitting that deal with breakdown as the initiation step of pitting exist (Ref 86, 87, 88, 89, 90). Mechanisms. Many theories or models have been proposed to describe the events leading to pitting or crevicecorrosion initiation. Successful models must, of course, explain the phenomenology of breakdown. The following phenomena are usually considered to be associated with chemical breakdown that leads to localized attack (Ref 86): • • •

•

A certain critical potential for breakdown must be exceeded. Damaging species (examples are chloride ions or the higher atomic weight halides) are needed to initiate and propagate breakdown. An induction time exists that starts with the initiation of the breakdown process by the introduction of conditions conducive to breakdown and ends with the completion of the process when breakdown commences. Highly localized sites exist at which breakdown occurs.

In order to develop a better understanding of breakdowns that can lead to an improved resistance to localized corrosion, three groups of models for passive-film breakdown have been proposed; these groups are described below. Adsorbed Ion Displacement Model (Ref 91, 92). The passive film is considered by this model to be an adsorbed oxygen film (probably a monolayer). Breakdown occurs when a more strongly adsorbing damaging anion, for example, a chloride ion, displaces the oxygen forming the passive film. After the chloride ion is adsorbed on the surface, the breakdown process is initiated because the bonding of the metal ions to the metal lattice is weakened. In ion migration or penetration models, damaging anions move through the passive film; the breakdown process is complete when an anion reaches the metal/film interface. All of these models consider the passive film to be three dimensional. They differ widely in their proposed mode of penetration. At one extreme is a model assuming the existence of pores in the passive film (Ref 93, 94). The other type of penetration models are those involving migration of the damaging anion through a lattice, via defects or via some sort of ion-exchange process. Ion migration in a lattice can occur in a variety of ways (Ref 95, 96, 97, 98). Some of these models postulate that the penetrating chloride ions occupy sites in the lattice, and recent x-ray absorption spectroscopic studies find that evidence for the existence of chloride in the lattice of the passive film (Ref 99, 100). Breakdown repair or film-tearing models involve many dynamic breakdown-repassivation events during which chemically or electrochemically induced mechanical disruption of the passive film is followed by repair of the break. This dynamic process will then lead to the breakdown of passivity (Ref 101, 102). More recent studies (Ref 99, 103) have shown that metastable pitting events—breakdown-repair processes that occur below the critical potential for breakdown, Ecrit—may be involved in breakdown. One of these studies (Ref 99) proposes a mechanism that describes the events of metastable pitting that lead to stable pit growth as follows: • •

Anion (e.g., chloride-ion) movement through the passive film at local sites under an electric field Formation of metal chloride at discrete sites at the passive-film/metal interface

• •

Initiation of pitting upon rupture of the film at metal-chloride sites Pit growth at exposed sites sustained when chloride ions under diffusion control can prevent repassivation

It has been pointed out (Ref 91), however, that stable pitting (occurs above Ecrit), rather than metastable pitting (observed below Ecrit), is from an engineering standpoint the real corrosion risk. Effect of Alloy Composition and Structure. Alloy composition has been found to affect breakdown phenomenologically by shifting Ecrit in the noble direction (Ref 104). This shift has been explained by the production of a passive film that is more difficult to penetrate because it provides fewer diffusion paths (Ref 47, 105), by alloying elements affecting repassivation kinetics (Ref 106), or by the formation of complexes with, for example, molybdenum, an alloy component that increases resistance to pitting by reducing the flux of cation vacancies in the passive film toward the film/metal interface and thereby increasing the induction time for breakdown (Ref 107). Another example has been found for amorphous and partially nanocrystalline alloys of aluminum that exhibit increased pit-growth potentials, reduced pit-propagation rates, and increased repassivation rates when compared to polycrystalline high-purity aluminum (Ref 108). This is an effect of both adding alloy elements and changing the alloy structure. Alloy structure determines the sites on a surface where the breakdown of the passive film is initiated. These sites have been shown to be related to the defect structure of the underlying metal (Ref 90, 91), with the density of sites in many instances depending on the crystallographic orientation of a particular grain (Ref 109). Another important factor leading to the production of breakdown sites is the presence of nonmetallic inclusions, especially the manganese sulfide inclusions found in stainless steels (Ref 90, 91). Finally, intermetallic phase particles acting as local cathodes (Ref 110, 111) as, for example, the FeAl3 in iron-contaminated aluminum alloys, raise the pH of the local environment and cause alkaline dissolution (Ref 91) of the matrix at its boundary with the particle and thereby initiate breakdown.

References cited in this section 47. T.P. Hoar, J. Electrochem. Soc., Vol 117, 1970, p 17C–22C 86. J. Kruger and K. Rhyne, Nucl. Chem. Waste Manage., Vol 3, 1982, p 205–227 87. J.R. Galvele, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 285–327 88. M. Janik-Czachor, J. Electrochem. Soc., Vol 128,


References 1. H.H. Uhlig, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 1–28 2. A.G. Revesz and J. Kruger, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 137

3. R.P. Frankenthal and J. Kruger, Ed., Passivity of Metals, Electrochemical Society, 1978 4. J. Kruger, Int. Mater. Rev., Vol 33, 1988, p 113–130 5. H. Hasegawa and T. Sugano, Ed., Passivation of Metals and Semiconductors, Part II, Passivity of Semiconductors, Pergamon Press, 1990 6. K.E. Heusler, Ed., Passivation of Metals and Semiconductors, Materials Science Forum, Vol 185–188, Trans Tech Publications, 1995 7. M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Passivity of Metals and Semiconductors, Proc. Vol 99–42, Electrochemical Society, 2001 8. C. Wagner, Corros. Sci., Vol 5, 1965, p 751–764 9. Definitions of Terms Relating to Corrosion and Corrosion Testing, G 13, Annual Book of ASTM Standards, ASTM, 1983 10. D. Peckner and I.M. Bernstein, Handbook of Stainless Steels, McGraw-Hill, 1977, Ch 24, p 24 11. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, 2nd ed., Pergamon Press, 1966 12. M.G. Fontana and N.D. Greene, Corrosion Engineering, McGraw-Hill, 1967, p 336–337 13. N. Sato, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 29– 58 14. M. Cohen, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 521–545 15. H.H. Uhlig, Z. Elektrochem., Vol 62, 1958, p 626–632 16. U.R. Evans, J. Chem. Soc., Vol 1927, 1927, p 1020–1040 17. J. Kruger, J. Electrochem. Soc., Vol 110, 1963, p 654–663 18. J. Kruger and J.P. Calvert, J. Electrochem. Soc., Vol 114, 1967, p 43–49 19. P.M.G. Draper, Corros. Sci., Vol 7, 1967, p 91–101 20. M. Nagayama and M. Cohen, J. Electrochem. Soc., Vol 109, 1962, p 781–790 21. K. Kubanov, R. Burstein, and A. Frumkin, Disc. Faraday Soc., Vol 1, 1947, p 259–269 22. R.P. Frankenthal, Electrochim. Acta, Vol 16, 1971, p 1845–1857 23. R.P. Frankenthal, J. Electrochem. Soc., Vol 114, 1967, p 542–547 24. K.E. Heusler, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 771–801 25. R.W. Revie, B.G. Baker, and J.O'M. Bockris, J. Electrochem. Soc., Vol 122, 1975, p 1460–1466 26. M.C Bloom and M. Goldenberg, Corros. Sci., Vol 5, 1965, p 623–630

27. B.D. Cahan and C.-T. Chen, J. Electrochem. Soc., Vol 129, 1982, p 921–925 28. K.S. Vetter, Z. Elektrochem., Vol 62, 1958, p 642–648 29. C.L. Foley, J. Kruger, and C.J. Bechtoldt, J. Electrochem. Soc., Vol 114, 1967, p 944–1001 30. H.T. Yolken, J. Kruger, and J.P. Calvert, Corros. Sci., Vol 8, 1968, p 103–108 31. K. Asami, K. Hashimoto, T. Musumoto, and S. Shimodaira, Corros. Sci., Vol 16, 1976, p 909–914 32. K. Kuroda, B.D. Cahan, Ch. Nazri, E. Yeager, and T.E. Mitchell, J. Electrochem. Soc., Vol 129, 1982, p 2163–2169 33. L.J. Jablonsky, M.P. Ryan, and H.S. Isaacs, J. Electrochem. Soc., Vol 145, 1998, p 1922–1932 34. W.E. O'Grady, J. Electrochem. Soc., Vol 127, 1980, p 555–563 35. G. Okamoto and T. Shibata, Nature, Vol 206, 1965, p 1350 36. G.G. Long, J. Kruger, D.R. Black, and M. Kuriyama, J. Electrochem. Soc., Vol 130, 1983, p 240–242 37. M. Kekar, J. Robinson, and A.J. Forty, Faraday Discuss. Chem. Soc., Vol 89, 1990, p 31–40 38. A.J. Davenport, and M. Sansone, J. Electrochem. Soc., Vol 142, 1995, p 727–730 39. J. Kruger, L.A. Krebs, G.G. Long, J.F. Anker, and C.F. Majkrzak, Passivation of Metals and Semiconductors, K.E. Heusler, Ed., Materials Science Forum, Vol 185–188, Trans Tech Publications, 1995, p 367–376 40. C.R. Clayton, K. Doss, and J.B. Warren, Passivity of Metals and Semiconductors, M. Froment, Ed., 1983, p 585–590 41. A.J. Davenport, H.S. Isaacs, J.A. Bardwell, B. MacDougall, G.S. Frankel, and A.G. Schrott, Corros. Sci., Vol 35, 1993, p 19–25 42. A.J. Davenport, M. Sansone, J.A. Bardwell, A.J. Andlykiewicz, M. Taube, and C.M. Vitus J. Electrochem. Soc., Vol 141, 1994, p L6–8 43. J. Eldrige, M.E. Kordesch, and R.W. Hoffman, J. Vac. Sci. Technol., Vol 20, 1982, p 934–938 44. G.G. Long, J. Kruger, D.R. Black, and M. Kuriyama, J. Electroanal. Chem., Vol 150, 1983, p 603–610 45. L.J. Oblonsky, A.J. Davenport, M.P. Ryan, and M.F. Toney, Passivity of Metals and Semiconductors, M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Proc. Vol 99-42, Electrochemical Society, 2001, p 173–177 46. A.J. Davenport, R.C. Newman, and P. Ernst, Passivity of Metals and Semiconductors, M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Proc. Vol 99-42, Electrochemical Society, 2001, p 65–71 47. T.P. Hoar, J. Electrochem. Soc., Vol 117, 1970, p 17C–22C 48. M.P. Ryan, R.C. Newman, and G.E. Thompson, Philos. Mag. B, Vol 70, 1994, p 241–251

49. M.P. Ryan, S. Fugimoto, G.E. Thompson, and R.C. Newman, Passivation of Metals and Semiconductors, K.E. Heusler, Ed., Materials Science Forum, Vol 185–188, Trans Tech Publications, 1995, p 233–240 50. G.G. Long, J. Kruger, M. Kuriyama, D.R. Black, E. Farabaugh, D.M. Saunders, and A.I. Goldman, Proc. Ninth Int. Congress on Metallic Corrosion, Proc. Vol 3, National Research Council, Ottawa, Canada, 1984, p 419–422 51. T. Shibata and Y.-C. Zhu, Corros. Sci., Vol 36, 1994, p 153–163 52. M. Pankuch, R. Bell, and C.A. Melendres, Electrochim. Acta, Vol 38, 1993, p 2777–2779 53. K.E. Healy and P. Ducheyne, Mater. Sci., Vol 4, 1993, p117–126 54. J.S.L. Leach and B.R. Pearson, Corros. Sci., Vol 28, 1988, p 43–56 55. P. Marcus and V. Maurice, Passivity of Metals and Semiconductors, M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Proc. Vol 99-42, Electrochemical Society, 2001, p 30–54 56. V. Vignal, J.M. Olive, and J.C. Roux, Passivity of Metals and Semiconductors, M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Proc. Vol 99-42, Electrochemical Society, 2001, p 208–213 57. W. Yang, D. Costa, and P. Marcus, J. Electrochem. Soc., Vol 141, 1994, p 2669–2676 58. K.J. Vetter, J. Electrochem. Soc., Vol 110, 1963, p 597–605 59. F.M. Delnick and N. Hackerman, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 116–133 60. E.K. Oshe, I.L. Rosenfeld, and V.C. Doizoskenko Dokl. Akad. Nauk SSSR, Vol 194, 1970, p 612–614 61. M. Bojinov, T. Laitinen, K. Mäkelä, T. Saario, T. Sirkiä, and G. Fabricius, Passivity of Metals and Semiconductors, M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Proc. Vol 99-42, Electrochemical Society, 2001, p 201–207 62. W. Schmickler, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 102–115 63. W. Schultze and U. Stimming, Z. Phys. Chem., NF, Vol 98, 1975, p 285–302 64. J. Ord and D.J. DeSmet, J. Electrochem. Soc., Vol 113, 1966, p 1258–1262 65. R.V. Moshtev, Electrochim. Acta., Vol 16, 1971, p 2039–2048 66. S.F. Bubar and D.A. Vermilyea, J. Electrochem. Soc., Vol 113, 1966, p 892–895 67. S.F. Bubar and D.A. Vermilyea, J. Electrochem. Soc., Vol 114, 1967, p 882–885 68. J.S.L. Leach and P. Neufeld, Corros. Sci., Vol 9, 1969, p 225–244 69. N. Sato and G. Okamoto, Comprehensive Treatise of Electrochemistry, J.O'M. Bockris et al., Ed., Vol 4, Plenum Press, 1981, p 193–306 70. U.F. Frank, Z. Naturforsch., Vol 4A, 1949, p 378–391

71. U. Ebersbach, K. Schwabe, and K. Ritter, Electrochim. Acta, Vol 12, 1967, p 927–938 72. A.T. Fromhold, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 59–81 73. N. Cabrera and N. Mott, Rep. Prog. Phys., Vol 12, 1949, p 163–184 74. J. Kruger and J.P. Calvert, J. Electrochem. Soc., Vol 114, 1967, p 43–49 75. A.T. Fromhold, Jr., and J. Kruger, J. Electrochem. Soc., Vol 120, 1973, p 722–729 76. C.Y. Chao, L.F. Lin, and D.D. Macdonald, J. Electrochem. Soc., Vol 128, 1981, p 1187–1194 77. Ya.M. Kolotyrkin and V.M. Knyazheva, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 678–698 78. N. Sato and M. Cohen, J. Electrochem. Soc., Vol 111, 1964, p 513–519 79. M. Sakashita and N. Sato, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 479–483 80. K.J. Vetter and F. Gorn, Electrochim. Acta, Vol 18, 1973, p 321–326 81. F. Flade, Z. Phys. Chem., Vol 76, 1911, p 513–559 82. H.H. Uhlig and P.F. King, J. Electrochem. Soc., Vol 106, 1959, p 1–7 83. A.J. Davenport, H.S. Isaacs, J.A. Bardwell, B. MacDougall, G.S. Frankel, and A.J. Schrott, Corros. Sci., Vol 35, 1993, p 19–25 84. L.J. Oblonsky, Passivity of Metals and Semiconductors, M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Proc. Vol 99-42, Electrochemical Society, 2001, p 253–257 85. J. Kruger, Proc. Corrosion and Corrosion Protection, R.P. Frankenthal and F. Mansfeld, Ed., Vol 81-8, Electrochemical Society, 1981, p 66–76 86. J. Kruger and K. Rhyne, Nucl. Chem. Waste Manage., Vol 3, 1982, p 205–227 87. J.R. Galvele, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 285–327 88. M. Janik-Czachor, J. Electrochem. Soc., Vol 128, 1981, p 513C–519C 89. Z. Szlarska-Smialowska, Pitting Corrosion of Metals, National Association of Corrosion Engineers, 1986 90. H. Böhni, Uhlig's Corrosion Handbook, R.W. Revie, Ed., 2nd ed., John Wiley, 2000, p 173–190 91. Ya.M. Kolotyrkin, J. Electrochem. Soc., Vol 108, 1961, p 209–216 92. H.P. Leckie and H.H. Uhlig, J. Electrochem. Soc., Vol 113, 1966, p 1262–1267 93. U.R. Evans, L.C. Bannister, and S.C. Britton, Proc. R. Soc., Vol 131A, 1931, p 367–376

94. J.A. Richardson and G.C. Wood, Corros. Sci., Vol 10, 1970, p 313–323 95. M.A. Heine, D.S. Keir, and M.I. Pryor, J. Electrochem. Soc., Vol 112, 1965, p 24–32 96. C.L. McBee and J. Kruger, Localized Corrosion, R.W. Staehle, B.F. Brown, J. Kruger, and A. Agrawal, Ed., National Association of Corrosion Engineers, 1974, p 252–260 97. L.F. Lin, C.Y. Chao, and D.D. Macdonald, J. Electrochem. Soc., Vol 128, 1981, p 1194–1198 98. K.E. Heusler and L. Fischer, Werkst. Korros., Vol 27, 1976, p 551–558 99. G.T. Burstein, G.O. Ilevbare, C. Liu, and R.M. Souto, Passivity of Metals and Semiconductors, M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Proc. Vol 99-42, Electrochemical Society, 2001, p 499–504 100. L.A. Krebs and J. Kruger, Passivity of Metals and Semiconductors, M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Proc. Vol 99-42, Electrochemical Society, 2001, p 561–566 101. J. Kruger, Passivity and Its Breakdown on Iron and Iron Base Alloys, R.W. Staehle and H. Okada, Ed., National Association of Corrosion Engineers, 1976, p 91–98 102.

N. Sato, Electrochim. Acta, Vol 16, 1971, p 1683–1692

103. G.S. Frankel, Passivity of Metals and Semiconductors, M.B. Ives, J.L. Luo, and J.R. Rodda, Ed., Proc. Vol 99-42, Electrochemical Society, 2001, p 445–476 104.

Ya.M. Kolotyrkin, Corrosion, Vol 19, 1963, p 261t–270t

105. C.L. McBee and J. Kruger, Passivity and Its Breakdown on Iron and Iron Base Alloys, R.W. Staehle and H. Okada, Ed., National Association of Corrosion Engineers, 1976, p 131–132 106.

T. Kodama and J.R. Ambrose, Corrosion, Vol 33, 1977, p 155–161

107.

M. Urguidi and D.D. Macdonald, J. Electrochem. Soc., Vol 132, 1985, p 555–558

108. J.E. Sweitzer, J.R. Scully, R.A. Bley, and J.W.P. Hsu, Electrochem. Solid-State Lett., Vol 2, 1999, p 267–270 109.

J. Kruger, J. Electrochem. Soc., Vol 106, 1959, p 736

110.

G.A.W. Murray, H.J. Lamb, and H.P. Godard, Brit. Corros. J., Vol 2, 1967, p 216–218

111. J.O. Park, C.H. Paik, and R.C. Alkire, Critical Factors in Localized Corrosion II, P.M. Natishan, R.J. Kelly, G.S. Frankel, and R.C. Newman, Ed., Electrochemical Society, 1995, p 218–225

J.R. Scully and R.G. Kelly, Methods for Determining Aqueous Corrosion Reaction Rates, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 68–86

Methods for Determining Aqueous Reaction Rates

Corrosion

John R. Scully and Robert G. Kelly, University of Virginia

Introduction CORROSION OF MATERIALS IN AQUEOUS SOLUTIONS is often thermodynamically possible but kinetically limited. Therefore, it is important to determine the rates of corrosion processes. Corrosion rate determination can serve many engineering and scientific purposes. For example, it can be used to: • • • • • • • • •

Screen available materials to find the most resistant material for a given application. Determine operating conditions where corrosion rates are low versus those where rates are high, by varying conditions. Determine probable service lifetimes of materials forming components, equipment, and processes. Evaluate new alloys or treatments or existing alloys in new environments. Evaluate lots, heats, or treatments of materials to ensure that specified quality is achieved before release, shipment, or acceptance. Evaluate environmental conditions such as new chemical species, inhibitors, or plant- operation conditions such as temperature excursions. Determine the most economical means of reducing corrosion through use of inhibitors, pretreatments, coatings, or cathodic protection. Determine the relative corrosivity of one environment compared to another. Study corrosion mechanisms.

Methods for determination of corrosion rates can be differentiated between those that measure the cumulative results of corrosion over some period of time and those that provide instantaneous rate information. Corrosion rates do not often increase monotonically with environmental conditions but exhibit sharp thresholds that distinguish regions of low corrosion rates from other regions where corrosion rates are dangerously high. It is sometimes of greater interest to define these thresholds than it is to determine the rates in the regions where corrosion rates are high. Examples of the latter are pitting or crevice corrosion where passive films are broken down and local corrosion rates can be extremely high. This article addresses electrochemical methods for instantaneous rate determination and threshold determination as well as nonelectrochemical methods that can determine incremental or cumulative rates of corrosion.

J.R. Scully and R.G. Kelly, Methods for Determining Aqueous Corrosion Reaction Rates, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 68–86 Methods for Determining Aqueous Corrosion Reaction Rates John R. Scully and Robert G. Kelly, University of Virginia

Electrochemical Methods Several textbooks and symposia proceedings cover the application of electrochemical methods that can be used for corrosion testing (Ref 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16). This article highlights many of the commonly used laboratory methods used to determine instantaneous corrosion rates as well as methods that characterize corrosion thresholds that signal conditions for high rates of corrosion. Techniques discussed include Tafel extrapolation, polarization resistance, electrochemical impedance spectroscopy, electrochemical noise resistance, use of rotating disks and cylinders to study aspects of corrosion affected by solution flow, polarization methods for assessing susceptibility to localized corrosion, scratch repassivation, potential-step repassivation, electrochemical noise applied to pitting, potentiodynamic repassivation techniques for assessing sensitization, and methods for determining underpaint corrosion rates associated with corrosion under organic coatings on metals.

References cited in this section 1. J.O'M. Bockris and A.K.N. Reddy, Modern Electrochemistry-2, Plenum Press, 1970 2. E. Gileadi, Electrode Kinetics for Chemists, Chemical Engineers and Materials Scientists, VCH Publishers, 1993 3. N.D. Tomashov, Theory of Corrosion and Protection of Metals, Macmillan Publishing, 1966 4. D.A. Jones, Principles and Prevention of Corrosion, 2nd ed., Macmillan Publishing, 1996 5. M.G. Fontana and N.D. Greene, Corrosion Engineering, McGraw-Hill, 1978 6. J. Newman, Electrochemical Systems, Prentice-Hall, 1973 7. H.H. Uhlig and R.W. Revie, Corrosion and Corrosion Control, John Wiley & Sons, 1985 8. J.O'M. Bockris, B.E. Conway, E. Yeager, and R.E. White, Ed., Electrochemical Materials Science, Vol 4, Comprehensive Treatise of Electrochemistry, Plenum Press, 1981 9. R. Baboian, Ed., Electrochemical Techniques for Corrosion, National Association of Corrosion Engineers, 1977 10. U. Bertocci and F. Mansfeld, Ed., Electrochemical Corrosion Testing, STP 727, American Society for Testing and Materials, 1979 11. Laboratory Corrosion Tests and Standards, STP 866, G. Haynes and R. Baboian, Ed., American Society for Testing and Materials, 1985 12. Corrosion Monitoring in Industrial Plants Using Nondestructive Testing and Electrochemical Methods, STP 908, G.C. Moran and P. Labine, Ed., American Society for Testing and Materials, 1986

13. R. Baboian, Ed., Electrochemical Techniques for Corrosion Engineers, National Association of Corrosion Engineers, 1986 14. R. Baboian, W.D. France, Jr., L.C. Rowe, and J.F. Rynewicz, Ed., Galvanic and Pitting CorrosionField and Laboratory Studies, STP 576, American Society for Testing and Materials, 1974 15. B. Poulson, Corros. Sci., Vol 23 (No. 4), 1983, p 391–430 16. D.D. MacDonald, Transient Techniques in Electrochemistry, Plenum Press, 1977


Electrode Reaction Thermodynamics and Kinetics in Corrosion Metallic corrosion is usually an electrochemical process. Electrochemical processes require anodes and cathodes in electrical contact as well as an ionic conduction path through an electrolyte. The electron flow between the anodic and cathodic areas quantifies the rates of the oxidation and reduction reactions. When anodes are physically separated from cathodes, the current can be readily measured by replacing direct electrical contact with a zero-resistance ammeter (ZRA), The conversion of the reaction rate per unit area, J, from units of mol/cm2 · s to electrical current density, (A/cm2), is accomplished with Faraday's constant, F, and knowledge of the number of electrons, n, transferred to complete the electrochemical reaction a single time: i = I/A = nFJ

(Eq 1)

where I is the current resulting from an electrochemical reaction over a known electrode surface area, n is the number of electrons transferred (equivalents/mol), F is Faraday's constant (96,484.6 Coulombs/equivalent), and A is the electrode surface area. Monitoring this current provides the means for assessing the kinetics of the corrosion process, not just the thermodynamic tendencies for corrosion or merely the mass of metal loss registered after the test. Typical corrosion processes occurring under freely corroding conditions involve at least one cathodic and one anodic reaction. The thermodynamics discussed previously dictate the circumstances where these reactions will proceed spontaneously. The current measured during a polarization experiment, iapp, involving a single- chargetransfer-controlled oxidation reaction and single charge-transfer-controlled reduction reaction is:

(Eq 2)

where icorr is the corrosion current density, E is the applied potential, R is the ideal gas constant, T is the temperature, and α is the transfer coefficient. The corrosion potential, Ecorr, is a kinetically and thermodynamically determined “mixed” potential given by the interception of the lines describing the total anodic and cathodic reaction rates. At Ecorr, iox = ired and icorr is described by the magnitude of iox at Ecorr as shown in Fig. 1. Either reaction rate may become mass-transport limited under certain circumstances, in which case Eq. 2 does not apply.

Fig. 1 Application of mixed-potential theory showing the electrochemical potential-current relationship for a corroding system consisting of a single charge-transfer-controlled cathodic reaction and chargetransfer-controlled anodic electrochemical reaction. βc and βa are Tafel slopes. Source: Ref 17 Obtaining Corrosion Rates from Electrochemical Kinetic Data. According to mixed- potential theory, any overall electrochemical reaction can be algebraically divided into half-cell oxidation and reduction reactions in which there can be no net electrical charge accumulation (Ref 17). For open-circuit corrosion in the absence of an applied potential, the oxidation of the metal and the reduction of some specie in solution occur simultaneously at the metal/electrolyte interface as described by Eq 2. Under these circumstances, the net measurable current density, iapp, is zero. However, a finite rate of corrosion defined by icorr occurs at local anodic sites on the metal surface as indicated in Fig. 1. When the corrosion potential, Ecorr, is located at a potential that is distinctly different from the reversible electrode potentials (Eredox) of either the corroding metal or the species in solution that is cathodically reduced, the oxidation of cathodic reactants or the reduction of any metallic ions in solution becomes negligible. Because the magnitude of iox at Ecorr is the quantity of interest in the corroding system, this parameter must be determined independently of the oxidation reaction rates of other adsorbed or dissolved reactants.

The information obtained in a polarization experiment is iapp as a function of the potential, E, as shown by the thick solid line in Fig. 1, where iapp = iox - ired. To obtain iapp as a function of E, the applied potential between the reference electrode and working electrode is controlled and scanned at constant rate (potentiodynamic), instantaneously increased a fixed amount, or stepped at various times (potentiostaircase) (Ref 18). The applied current, Iapp, is measured and normalized with respect to the surface area (i.e., iapp = Iapp/A). Conversely, iapp can be supplied between the working and counterelectrodes under galvanostatic or galvanostaircase control, and the resulting potential between the working and reference electrodes is monitored. Several ASTM standards discuss methods for performing these experiments (Ref 19, Ref 20, 21). Several approaches are available to determine icorr from such experimental information. These methods are discussed in the sections that follow. Conversion of Corrosion Rates to Mass Loss and Penetration Rate. Determination of Iox at open-circuit potential or other potential of interest, where Iox = iox × area, over a known period of time leads to direct determination of the mass loss: (Eq 3) where M is the mass loss (g), Ioxt is the product of current and time (coulombs), and AW is the atomic weight of the electroactive species (g/ mol). This relation is known as Faraday's First Law. Rearrangement of Eq 3 leads to a straightforward determination of the corrosion penetration rate (applicable only when iox, μA/cm2, is uniformly distributed over the entire wetted surface area or where the localized actively corroding area, A, is known), as follows: (Eq 4) where CR is the corrosion penetration rate (in mm/yr, EW is the equivalent weight (considered dimensionless in this calculation because the units of equivalents/mol are included in K1), and ρ is the metal or alloy density, g/cm3. The constant K1 = 3.27 × 10-3 when iox is expressed as μA/cm2. K1 has units of mm · g/ μA · cm · yr when CR is desired in mm/yr. ASTM G 102 gives other values for this constant when CR is expressed in other units (Ref 22). Expressions are also available to calculate mass- loss rate per unit area, MR, from knowledge of electrochemical corrosion rate. For instance: (Eq 5)

MR = K2iox(EW) -3

2

2

where K2 = 8.954 × 10 g · cm /μA · m · d when MR is expressed as g/m2 · day. For alloys, the equivalent weight EW should be calculated as outlined in ASTM G 102 (Ref 22): (Eq 6) where fi is the mass fraction of the ith component of the alloy (-), AWi is the atomic weight of the ith component element (g/mol), ni is the number of electrons transferred or lost when oxidizing the ith component element under the conditions of the corrosion process (equivalents/mol), and i is the number of component elements in the alloy. n is usually equal to the stable valence of the elements oxidized from the metallic state or must be determined from either a Pourbaix (potential- pH) diagram or experimentally from an analysis of the corroding solution. This expression assumes that all the component elements oxidize when the alloy corrodes and that they are all oxidizing at essentially a uniform rate. In some situations, these assumptions are not valid; in these cases, the calculated corrosion rate will be in error. For example, if an alloy is composed of two or more phases and one phase preferentially corrodes, the calculation must take this into consideration.

References cited in this section 17. C. Wagner and W. Traud, Z. Electrochem., Vol 44, 1938, p 391

18. W.D. France, Jr., Controlled Potential Corrosion Tests: Their Application and Limitations, Mater. Res. Standard., Vol 9 (No. 8), 1969, p 21 19. “Standard Practice for Conventions Applicable to Electrochemical Measurements in Corrosion Testing,” G 3, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 20. “Standard Reference Test Method for Making Potentiostatic and Potentiodynamic Anodic Polarization Measurements,” G 5, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 21. “Practice for Conducting Potentiodynamic Polarization Resistance Measurements,” G 59, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 22. “Practice for Calculation of Corrosion Rates and Related Information from Electrochemical Measurements,” G 102, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM


Electrochemical Methods for the Study of Uniform Corrosion: Polarization Methods One method for examining the corrosion behavior of a metal is to determine the E-iapp relationship by conducting a polarization experiment. The following E-iapp relationship is often experimentally observed between applied current and potential. The expression is applicable to charge-transfer-controlled corrosion processes regardless of the exact number of charge-transfer-controlled reactions or reaction steps. It provides the basis for the electrochemical polarization technique (Ref 1, 2, 17, 18, 23):

(Eq 7)

where C is the interfacial capacitance associated with the electrochemical double layer (Ref 1, 2), and βa and βc are the apparent anodic and cathodic Tafel slopes (∂E/∂ log iapp) = 2.303RT/αF given by the slopes of the polarization curves in the anodic and cathodic Tafel regimes, respectively, and ∂E/∂t is the time rate of change in applied potential; that is, the voltage scan rate. The second term of the expression (C∂E/∂t) approaches 0 at low voltage scan rates dE/dt. This is desirable since the goal is to obtain icorr at Ecorr. Note that iapp becomes approximately equal to either iox or ired at large η, where η = (E - Ecorr). At very large anodic or cathodic overpotentials, Eq 7 can be rearranged in the form of the Tafel expression (Ref 1, 2, 3, 5, 7, 19, 23): (Eq 8a)

(Eq 8b) where ηa and ηc are the anodic and cathodic overpotentials, respectively. Equations 8a and (b) are only strictly valid for a single anodic or cathodic reaction, respectively, although approximately linear behavior can be observed despite several reactions. The linear portion of the solid line describing E-iapp in Fig. 1 above the Ecorr describes a single anodic reaction that can be described by Eq 8a. Tafel Extrapolation to Determine Corrosion Rate. In cases where Eq 8a and (b) are valid, η can be plotted versus log(iapp) over a sufficient range to obtain a linear relationship between the logarithm of current density and potential (Ref 1, 5, 7, 23), icorr is determined from extrapolation of iapp from either the anodic or the cathodic Tafel region to the open-circuit or corrosion potential (i.e., zero overpotential). Figure 1 illustrates this method. The method is potentially damaging to the corroding metal since a large overpotential must be applied. This is particularly true in the case of anodic polarization, in which the surface is changing because of corrosion and/or passivation of the metal. No ASTM standard currently exists for this method. Complications with the Polarization Method Involving Solution Resistance. Tafel extrapolation as well as other polarization techniques can be complicated by several factors. One such factor, ohmic resistance, arises from the resistivity of the solution, cell geometry, location of the reference electrode, and magnitude of applied current (Ref 2, 6). Ohmic resistance can contribute a voltage error to the measured potential. This error is summed algebraically with the true interfacial overpotential across the electrochemical interface (the potential usually sought in electrochemical measurements) measured with respect to a reference electrode. Placement of the reference electrode near the working electrode with a Luggin-Haber capillary is used to minimize the solution resistance error, which can be estimated from the product of the applied current density, the solution resistivity, and the perpendicular distance from the Luggin probe to the specimen surface in a planar electrode geometry. The error contributes to the measured overpotential: ηapp = ηtrue interfacial + iappRs

(Eq 9)

2

Rs(Ω · cm ) is the uncompensated solution resistance between the working electrode and the position where the reference electrode senses the potential in solution (at the tip of the Luggin- Haber capillary). Thus, ηapp > ηtrue at a high anodic or cathodic applied current density, Eapp > Etrue and the Tafel slope desired from a fit of Eq 7 or 8a to E versus iapp data is not obtained. When the dominant term in Eq 9 is the second term, a linear relationship between η and iapp is obtained instead of the semilogarithmic relationship discussed previously. The true scan rate in the potentiodynamic technique may also be altered by ohmic resistance since the applied potential that contains the ohmic component is controlled during the scan, not the true overpotential. Several excellent reviews are available on the subject of the voltage error introduced from solution resistance (Ref 24, 25, 26, 27). Complications Involving Concentration Polarization Effects. The Tafel relationship established through Eq 7 and 8a is dependent on pure activation control, or charge-transfer control. An additional consideration involves the concept of concentration polarization. In this case, the reaction rate is fast enough that the reacting specie is depleted (reduction reaction) or concentrated (oxidation) at the reacting surface. In order to maintain the reaction rate, diffusion through the electrolyte becomes the kinetic limitation. The reaction becomes diffusion controlled at the limiting current density, iL. The deviation from activation control in the case of a cathodic reaction can result in an additional overpotential known as a mass-transport overpotential, ηconc. This overpotential is described by (Ref 4, 5): (Eq 10) where iL is the mass transfer limiting current density defined by Fick's first law at steady state. As iapp approaches iL, the concentration overpotential, ηconc, becomes very large. The cathodic reaction may be under mixed charge-transfer/mass-transport or mass-transport control for many corrosion situations, particularly if the cathodic reaction is O2 reduction (Ref 5). The cathodic polarization behavior associated with “mixed” charge-transfer/mass-transfer control can be described mathematically by the algebraic sum of Eq 8a and 10. Tafel extrapolation of cathodic data becomes difficult under these conditions because the Tafel region is not extensive.

For a completely mass-transport-limited corrosion process, the concentration of the cathodic reacting specie Cb approaches 0 at the electrode interface and icorr = iL (Fig. 2) (Ref 2). The diffusional boundary layer thickness, δ, is decreased by increasing solution stirring or rotation rate, Ω(rad/s), in the case of a rotating cylinder or disk electrode (Ref 1). The limiting current density iL is a linear function of concentration gradient (Ref 15). The concentration gradient (Cb/δ) increases as a function of Ω0.5 or Ω0.7 for the rotating disk and cylinder, respectively (Ref 2, 15). For a mass-transport-controlled cathodic reaction, icorr is increased with flow as shown in Fig. 2. The governing equation for the rotating cylinder electrode is (Ref 15): iL = 0.079nFCbD0.64ν-0.34Ω0.7r0.4

(Eq 11)

and for the rotating-disk electrode is (Ref 2, 15): iL = 0.621nFCbD0.67ν-0.167Ω0.5

(Eq 12)

where Cb is the reacting species concentration in the bulk solution (moles/cm3), D is the diffusion coefficient for the reacting specie (cm2/s), ν is the kinematic viscosity of the solution (cm2/s), and r is the cylinder or disk radius (cm).

Fig. 2 Application of mixed-potential theory showing the electrochemical potential-current relationship for a corroding system consisting of a mass-transport-controlled cathodic reaction and a charge-transfercontrolled anodic reaction. As the fluid velocity increases from 1 to 4, the corrosion rate increases from A to D. Corrosion engineers often favor the rotating cylinder to simulate flow in turbulent piping systems since this flow regime is readily obtained (Ref 15). In contrast, the rotating-disk electrode operates in the laminar flow regime even at high rotation rates and does not accurately represent many corrosion situations (Ref 15). No ASTM methods currently exist to examine mass-transport-controlled corrosion kinetics.

References cited in this section 1. J.O'M. Bockris and A.K.N. Reddy, Modern Electrochemistry-2, Plenum Press, 1970 2. E. Gileadi, Electrode Kinetics for Chemists, Chemical Engineers and Materials Scientists, VCH Publishers, 1993 3. N.D. Tomashov, Theory of Corrosion and Protection of Metals, Macmillan Publishing, 1966 4. D.A. Jones, Principles and Prevention of Corrosion, 2nd ed., Macmillan Publishing, 1996 5. M.G. Fontana and N.D. Greene, Corrosion Engineering, McGraw-Hill, 1978 6. J. Newman, Electrochemical Systems, Prentice-Hall, 1973

7. H.H. Uhlig and R.W. Revie, Corrosion and Corrosion Control, John Wiley & Sons, 1985 15. B. Poulson, Corros. Sci., Vol 23 (No. 4), 1983, p 391–430 17. C. Wagner and W. Traud, Z. Electrochem., Vol 44, 1938, p 391 18. W.D. France, Jr., Controlled Potential Corrosion Tests: Their Application and Limitations, Mater. Res. Standard., Vol 9 (No. 8), 1969, p 21 19. “Standard Practice for Conventions Applicable to Electrochemical Measurements in Corrosion Testing,” G 3, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 23. Z. Tafel, Phys. Chem., Vol 50, 1904, p 641 24. D. Britz, J. Electroanal. Chem., Vol 88, 1978, p 309 25. M. Hayes and J. Kuhn, J. Power Sources, Vol 2, 1977–1978, p 121 26. F. Mansfeld, Corrosion, Vol 38 (No. 10), 1982, p 556 27. L.L. Scribner and S.R. Taylor, Ed., The Measurement and Correction of Electrolyte Resistance in Electrochemical Tests, STP 1056, American Society for Testing and Materials, 1990


Polarization Resistance Methods Stern and Geary simplified the kinetic expression describing charge-transfer-controlled reaction kinetics (Eq 7) for the case of small overpotentials with respect to Ecorr (Ref 28, 29). Equation 7 can be linearized when η/β < 0.1. This simplified relationship has the following form if the second term describing the capacitive current, C(∂E/∂t), is negligible: (Eq 13) rearranging: (Eq 14) where Rp is the polarization resistance (Ω · cm2) given by ∂E/∂i at t = ∞ and ΔE = 0. iapp is often approximately linear with potential within ±5 to 10 mV of Ecorr as shown for AISI 430 stainless steel in H2SO4 (Fig. 3). The slope of this plot, ΔE/Δi, when determined at Ecorr as shown in Fig. 3 defines the polarization resistance, which is inversely proportional to corrosion rate (Ref 30, 31). The surface area of the working electrode must be known. Knowledge of Rp, βa, and βc permit direct determination of the corrosion rate

at any instant in time using Eq 14 (Ref 28, 29, 31, 32). The ASTM standards G 59 (Ref 21) and G 96 (Ref 33) describe standard procedures for conducting polarization resistance measurements.

Fig. 3 ASTM G 59 polarization curves for polarization resistance measurements based on the results from eight independent laboratories for type 430 stainless steel in 1 N H2SO4. Curve 1 is the mean result, with curves 2 and 3 showing the 95% confidence limits. Source: Ref 21 This review focuses on corroding systems with different cathodic and anodic half-cell reactions. However, the concept of polarization resistance applies equally well to reduction-oxidation systems involving a single halfcell reaction. Here, the exchange current density, i0, may be calculated from the polarization resistance, where R is the ideal gas constant, F is Faraday's constant, T is the temperature, and αa and αc are the anodic and cathodic multistep electron-transfer coefficients, respectively, for the reduction-oxidation process.

Other techniques exploit nonlinearity at larger overpotentials that invalidate the approximation given by Eq 13 (Ref 34). The Oldham-Mansfeld method calculates icorr from nonlinear E versus iapp data obtained usually within ±30 mV of Ecorr using tangent slopes and intercepts to calculate corrosion rate without high overpotential determination of βa and βc (Ref 35). Computerized curve fitting can exploit nonlinearity to calculate βa, and βc from low overpotential data, avoiding the destructive nature of large overpotentials. The Mansfeld technique substitutes Eq 14 into Eq 7 eliminating icorr (Ref 36). βa and βc are determined from the best fit of the resulting expression containing βa and βc as unknowns to a nonlinear plot of η versus 2.3iappRp. η versus iapp data are obtained within ±30 mV of Ecorr. Rp is determined from the linear slope of E versus iapp within ±5 mV of Ecorr. icorr is subsequently determined from Eq 13 using the computed values of Rp, βa, and βc. Complications with polarization resistance measurements, and possible remedies are reported in the literature (Ref 34, 35, 36, 37, 38, 39, 40). Three of the most common errors involve: (a) invalidation of the results through oxidation of some other electroactive species besides the corroding metal in question, (b) a change in the open-circuit or corrosion potential during the time taken to perform the measurement, and (c) use of a large η, invalidating the assumption of a linear relationship between iapp and E required by Eq 13 and 14. Another source of error involves cases in which both the anodic and cathodic reactions are not charge-transfercontrolled processes, as required for the derivation of Eq 13. Modifications to Eq 7 exist for cases in which pure activation control is not maintained, such as in the case of partial diffusion control or passivation (Ref 41). Other researchers have attempted to calibrate the polarization resistance method with gravimetrically determined mass loss (Ref 42). In fact, polarization resistance data for a number of alloy- electrolyte systems have been compared to the observed average corrosion currents determined from mass loss via Faraday's law (Ref 32). A linear correspondence was obtained over six orders of magnitude in corrosion rates. Two other frequently encountered complications are the need to correct polarization data for errors that arise from the contribution of solution resistance, Rs, and the addition of capacitive current, CdE/dt, which occurs with increasing scan rate (Ref 16). Capacitive current gives rise to hysteresis in current-potential plots (Ref 43). Solution resistance contributes to a voltage error as discussed previously, as well as a scan-rate error. Since the applied potential is increased by an ohmic voltage component, an apparent value of polarization resistance is obtained that overestimates Rp by an amount equal to Rs. Consequently, the corrosion rate is underestimated. Hysteresis in the current density-applied potential plot is brought about for combinations of high voltage scan rate and large interfacial capacitances as well as large polarization resistances. Attempting to determine Rp at too fast of a scan rate underestimates its true value, leading to an overestimation of corrosion rate. This error can be minimized by determining the polarization resistance at a slow scan rate or by extrapolating the results at several slow scan rates to zero scan rate. Alternatively, one may take two or more current-density measurements from potentiostatic data after long time periods near Ecorr to minimize fast-scan-rate effects. These complications and others have been reviewed elsewhere (Ref 44). Many treatments of this subject have used an electrical equivalent circuit model to simulate the corroding metal/electrolyte interface (Ref 1, 44, 45). The simplest form of such a model is shown in Fig. 4. The three parameters discussed previously (Rp, Rs, and C) that approximate a corroding electrochemical interface are shown. The algebraic sum of Rs and Rp is measured when a direct-current measurement is performed. The impedance associated with a capacitor approaches infinity as the voltage scan rate approaches 0, and parallel circuit elements are always dominated by the element with the smallest impedance. Therefore, the sum of Rs and Rp is measured. The true corrosion rate will be underestimated when Rs is appreciable. Conversely, any experiment conducted at a fast voltage scan rate causes the algebraic sum of the ohmic resistance and the resultant impedance of the parallel resistive-capacitive network to be measured. This value will be lower than the sum of Rp and Rs determined at an infinitely slow scan rate, as current leaks through the parallel capacitive element at higher scan rate due to its low impedance at high frequency. This will usually result in an overestimation of the true corrosion rate. These complications can be overcome by using the electrochemical impedance method.

Fig. 4 Electrical equivalent circuit model simulating a simple corroding metal/electrolyte interface. See also Fig. 5. Rs is the solution resistance. Rp is the polarization resistance. C is the double-layer capacitance.

References cited in this section 1. J.O'M. Bockris and A.K.N. Reddy, Modern Electrochemistry-2, Plenum Press, 1970 16. D.D. MacDonald, Transient Techniques in Electrochemistry, Plenum Press, 1977 21. “Practice for Conducting Potentiodynamic Polarization Resistance Measurements,” G 59, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 28. M. Stern and A.L. Geary, J. Electrochem. Soc., Vol 104, 1957, p 56 29. S. Evans and E.L. Koehler, J. Electrochem. Soc., Vol 108, 1961, p 509 30. M. Stern and R.M. Roth, J. Electrochem. Soc., Vol l04, 1957, p 390 31. M. Stern and A.L. Geary, J. Electrochem. Soc., Vol 105, 1958, p 638 32. M. Stern and E.D. Weisert, Experimental Observations on the Relation between Polarization Resistance and Corrosion Rate, Proc. ASTM, Vol 59, 1959, p 1280 33. “Standard Guide for On-Line Monitoring of Corrosion in Plant Equipment (Electrical and Electrochemical Methods),” G 96, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 34. S. Barnartt, Electrochim. Acta, Vol 15, 1970, p 1313 35. K.B. Oldham and F. Mansfeld, Corros. Sci., Vol 13, 1973, p 813 36. F. Mansfeld, J. Electrochem. Soc., Vol 120, 1973, p 515 37. F. Mansfeld and M. Kendig, Corrosion, Vol 37 (No. 9), 1981, p 556 38. R. Bandy and D.A. Jones, Corrosion, Vol 32, 1976, p 126 39. M.J. Danielson, Corrosion, Vol 36 (No. 4), 1980, p 174 40. J.C. Reeve and G. Bech-Nielsen, Corros. Sci., Vol 13, 1973, p 351 41. I. Epelboin, C. Gabrielli, M. Keddam, and H. Takenouti, Electrochemical Corrosion Testing, STP 727, F. Mansfeld and U. Bertocci, Ed., American Society for Testing and Materials, 1981, p 150

42. A.C. Makrides, Corrosion, Vol 29 (No. 9), 1973, p 148 43. D.D. MacDonald and M.C.H. McKubre, Electrochemical Corrosion Testing, STP 727, F. Mansfeld and U. Bertocci, Ed., American Society for Testing and Materials, 1981, p 110 44. J.R. Scully, Corrosion, Vol 56 (No. 2), 2000, p 199–218 45. A.J. Bard and L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications, 2nd ed., John Wiley & Sons, 2001


Electrochemical Impedance Methods One approach for determining the polarization resistance of a metal involves the electrochemical impedance (sometimes known as alternating current, or ac, impedance) method (Ref 2, 43, 44, 45, 46, 47). ASTM G 106, “Practice for Verification of Algorithm and Equipment for Electrochemical Impedance Measurements,” contains an appendix reviewing the technique. In this technique, typically a small-amplitude sinusoidal potential perturbation is applied to the working electrode at a number of discrete frequencies, ω. ω is the angular velocity in rad/s, ω = 2πf, where f is frequency in Hz. At each one of the discrete frequencies, the resulting current waveform will exhibit a sinusoidal response that is out of phase with the applied potential signal by a certain amount (Φ) and has a current amplitude that is inversely proportional to the impedance of the interface. The electrochemical impedance, Z(ω), is the frequency-dependent proportionality factor that acts as a transfer function by establishing a relationship between the excitation voltage signal and the current response of the system: Z(ω) = V(ω)/i(ω)

(Eq 15)

where V is the time varying voltage across the circuit [V = V0 sin(ωt)], i is the time varying current density through the circuit [i = i0 sin (ωt + Φ)], Z(ω) is the impedance (Ω · cm2), and t is time (s), Z(ω) is a complex-valued vector quantity with real and imaginary components whose values are frequencydependent: Z(ω) = Z′(ω) + jZ″(ω)

(Eq 16)

where Z′(ω) is the real component of impedance [Z′(ω) = |Z(ω)| cos (Φ)], Z″(ω) is the imaginary component of impedance [Z″(ω) = |Z(ω)| sin (Φ)], j is the imaginary number , |Z(ω)| is the impedance magnitude, where |Z(ω)| = (Z′(ω)2 + Z″(ω)2)1/2. The electrochemical impedance is a fundamental characteristic of the electrochemical system it describes. Knowledge of the frequency dependence of impedance for a corroding system allows a determination of an appropriate equivalent electrical circuit describing that system. Table 1 shows the transfer functions for resistors, capacitors, and inductors.

Table 1 Linear circuit elements commonly used in electrochemical impedance Circuit element Resistor Capacitor Inductor

Impedance(a) Z(ω) = R Z(ω) = -j/ωC Z(ω) = jωL

(a) ω is the frequency; j = Figure 4 illustrates the equivalent electrical circuit model for a simple case of an actively corroding metal. The impedance for that system can be described by: (Eq 17) where ω is the frequency of the applied signal where ω = 2πf(rad/s), f is the frequency of the applied signal [Hz(cycles/s)], C is the interfacial capacitance (F/cm2). Rs is the solution resistance, and Rp is the polarization resistance. The Bode magnitude and phase information of Fig. 5 and Eq 17 show that at very low frequencies: Zω→0(ω) = Rs + Rp

(Eq 18)

while at very high frequencies: Zω→∞(ω) = Rs

(Eq 19)

Fig. 5 Bode phase angle and magnitude plots demonstrating the frequency dependence of electrochemical impedance for the circuit model shown in Fig. 4 Determination of Rp is attainable in media of high resistivity because Rp can be mathematically separated from Rs by taking the difference between Z(ω) obtained at low and high ω(Rp = Z(ω) → 0 - Z(ω) → ∞).

Determination of the corrosion rate using Eq 13 also requires knowledge of βa, βc and electrode area, A, which are not obtained in the impedance experiment (Ref 48). The minimum applied frequency required to obtain Rs + Rp can be approximated by: (Eq 20) where fbp is the lower breakpoint frequency (Hz) is approximated by the point on the log|Z(ω)| versus log f plot where the low-frequency plateau dominated by Rs + Rp and the slope -1 region dominated by capacitance produce equal values of Z(ω), and fmin is the minimum test frequency (Hz) required according to Eq 20. Since the magnitudes of C, Rs, and Rp are not known explicitly a priori, prudence dictates that fmin be selected as 0.1 to 0.5 of the estimated fbp. Thus, large values of C, Rs, or Rp dictate that a low fmin is required to accurately obtain Rp + Rs at Z(ω) → 0. One mHz is typically chosen as a reasonable initial choice of fmin, but it is obvious from Eq 20 that either a lower frequency may be required or a higher frequency permitted, depending on circumstances. Either the anodic or cathodic half-cell reaction can become mass-transport limited and restrict the rate of corrosion at Ecorr. The presence of diffusion-controlled corrosion processes does not invalidate the electrochemical impedance method but does require extra precaution and a modification to the circuit model of Fig. 4. In this case, the finite diffusional impedance is added in series with the usual charge-transfer parallel resistance shown in Fig. 4. The transfer function for the frequency-dependent finite diffusional impedance, ZD(ω), has been described (Ref 49): (Eq 21) where s = where leff is the actual finite diffusion length and D is the diffusivity of the diffusing species that limits the interfacial reaction. The value of ZD(ω) approaches the real component of diffusional resistance, RD, as ω → 0. The frequency required to obtain RD depends on the value of s. The larger the value of s, such as when leff is large or D is small, the lower the frequency required. Rp, defined as [∂E/∂iapp] as ω → 0, is the sum of the charge-transfer-controlled, Rct and diffusion-controlled, RD, contributions to the polarization resistance assuming that RD + Rct » Rs; that is: Rp = Rct + RD

(Eq 22)

A very low frequency or scan rate may be required to obtain Rp defined by Eq 22 under circumstances where reactions are mass-transport limited as indicated by Eq 21. For instance, in the case where leff = 0.1 cm and D = 10-5 cm2/ s a frequency below 0.1 mHz is required to obtain Rp from |Z(ω)| at the zero-frequency limit. Hence, a common experimental problem in the case of diffusion-controlled electrochemical reactions is that extremely low frequencies (or scan rates) are required to complete the measurement of Rp. In the case where Rp is dominated by contributions from mass transport such that Eq 22 applies, the Stern approximation of Eq 13 and 14 must be modified to account for a Tafel slope for either the anodic or cathodic reaction under diffusioncontrolled conditions (i.e., βa or βc = ∞). In fact, Eq 7 becomes invalid. Similarly, a frequency above fmax must be applied to obtain Rs: (Eq 23) where fmax is the frequency required such that Z(ω) is dominated by Rs. Typically, f must be in the kHz range to determine Rs. These issues equally plague time as well as frequency domain methods for obtaining Rp since in the time domain measurement, the triangle waveform is simply the Fourier synthesis of a series of sinusoidal signal functions. The capacitance is also determined from the impedance technique. In many corroding systems, an interfacial capacitance associated with the electrified double-layer scales linearly with the true electrochemical surface area. An electrochemically based estimate of the surface area may be obtained if the area-specific capacitance is known or determined from a plot of C versus surface area (Ref 48). A common issue in the use of impedance-derived capacitance concerns the use of constant- phase elements (CPE). The impedance associated with a CPE has been given by (Ref 50):

(Eq 24) Y0 is the constant-phase-element parameter. (In the case of the impedance due to capacitance, the value of the capacitance C, is the parameter used; that is, Y0 = C, in this expression, and n = 1). Y0 and n are usually assumed to be frequency- independent parameters. The units for Y0 are sn/ Ω, while those for capacitance (C) are s/Ω. Hence, in the case of an ideal capacitor or resistor then n = 1 or 0, respectively, and either the magnitude of Y0 equals the magnitude of C with the dimensions s/Ω or 1/Y0 = R(Ω). In the case of n = 0.5, an infinite diffusional impedance best describes the constant-phase element. At issue is the task of extracting physically meaningful parameters conveyed by the capacitance in the case of impedance data that are best represented in an electrical circuit model by a constant-phase element. Examples of such parameters extracted from ideal interfacial capacitance include electrode area in the case of a double- layer capacitance, surface coverage in the case of an adsorption pseudocapacitance, and dielectric constant or dielectric layer thickness in the case of a coating or oxide with dielectric properties. Other parameters may be extracted from capacitance associated with solid-state impedance experiments, but these are beyond the scope of this article. In the case where n = 0.8 to 0.99, the CPE is often treated as a nonideal capacitance value and attempts are made to extract physically meaningful parameters from the CPE data. One equation proposed to convert Y0 into C is (Ref 51): (Eq 25) is the frequency where the imaginary component of impedance, Z″, is maximized. This frequency where is independent of n. In an earlier approach, C = Y0(ω)n-1/sin(nπ/2) was suggested as a method for extracting capacitance, C, from Y0 where ω was taken as the frequency where the phase angle was maximized (Ref 52). This method has the disadvantage that the exact value of ω associated with the phase angle maximum changes with the value of n.

References cited in this section 2. E. Gileadi, Electrode Kinetics for Chemists, Chemical Engineers and Materials Scientists, VCH Publishers, 1993 43. D.D. MacDonald and M.C.H. McKubre, Electrochemical Corrosion Testing, STP 727, F. Mansfeld and U. Bertocci, Ed., American Society for Testing and Materials, 1981, p 110 44. J.R. Scully, Corrosion, Vol 56 (No. 2), 2000, p 199–218 45. A.J. Bard and L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications, 2nd ed., John Wiley & Sons, 2001 46. F. Mansfeld, Corrosion, Vol 36 (No. 5), 1981, p 301 47. “Practice for Verification of Algorithm and Equipment for Electrochemical Impedance Measurements,” G 106, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 48. J.N. Murray, P.J. Moran, and E. Gileadi, Corrosion, Vol 44 (No. 8), 1988, p 533 49. D.R. Franceschetti and J.R. Macdonald, J. Electroanal. Chem., Vol 101, 1979, p 307 50. B.A. Boukamp, Solid State Ionics, Vol 20, 1986, p 31–44 51. C.H. Hsu and F. Mansfeld, Corrosion, Vol 57, 2001, p 747 52. S.F. Mertens, C. Xhoffer, B.C. De Cooman, and E. Temmerman, Corrosion, Vol 53, 1997, p 381


Frequency Modulation Methods Both harmonic and electrochemical frequency modulation (EFM) methods take advantage of nonlinearity in the voltage-current (E-I) response of electrochemical interfaces to determine corrosion rate (Ref 53, 54, 55, 56). A special application of harmonic methods involves harmonic impedance spectroscopy (Ref 57). The EFM method uses one or more ac voltage perturbations in order to extract corrosion rate. The electrochemical frequency modulation method has recently been reviewed (Ref 58). In the most often used EFM method, a potential perturbation by two sine waves of different frequencies is applied across a corroding metal interface. The E-I behavior of corroding interfaces is typically nonlinear such that a potential perturbation in the form of a sine wave at one or more frequencies can result in a current response at the same and at other frequencies. The result of such a potential perturbation is various ac current responses at various frequencies such as zero, harmonic, and intermodulation. The magnitude of these current responses can be used to extract information on the corrosion rate of the electrochemical interface or conversely the reduction-oxidation rate of an interface dominated by redox reactions as well as the Tafel parameters. This is an advantage over linear polarization resistance and EIS methods, which can provide the Z(ω) and the polarization resistance of the corroding interface at ω = 0, but do not uniquely determine Tafel parameters in the same set of data. Separate experiments must be used to define Tafel parameters. A special extension of the method involves harmonic impedance spectroscopy where the harmonic currents are converted to harmonic impedance values at various frequencies through knowledge of the magnitude of the ac perturbation. In the EFM method, an ac voltage perturbation is applied at two frequencies, ω1 and ω2. As an example the voltage perturbation could be given as: η = V0[sin(ω1t) + sin(ω2t)]

(Eq 26)

where η = E - Ecorr and V0 is the magnitude of the voltage amplitude applied. Harmonic current responses occur at ω1, 2ω1, 3ω1, as well as at ω2, 2ω2, and 3ω2, and so forth. Additionally, a current response can be seen at various intermodulation frequencies such as 2ω1 ± ω2 and 2ω2 ± ω1. Consider the application of this method to a charge-transfer-controlled corrosion process with an E-I response that behaves according to Eq 7. Under the assumption that ω2 > ω1 and βa < βc, the corrosion current density, icorr, and Tafel parameters, ba (where ba = βa/ln 10) and bc, can be determined from the equations summarized in Table 2. The current components at the angular frequency ω1 or ω2 can be measured at ω1 or ω2. The intermodulation components ω1 ± ω2 can be determined from the signal response at ω1 + ω2 or ω1 - ω2 and so forth. The method is one of the few that enable extraction of corrosion rate and Tafel parameters directly from a single measurement (also see Ref 34 and 35). Currently, there are no ASTM standards for this technique. Table 2 Governing equations for extraction of icorr, as well as ba and bc from harmonic and intermodulation frequency data Reaction type Chargetransfercontrolled Tafel behavior

Governing equation

Determination of icorr

Determination parameters

of

Tafel

Passive or anodic masstransport control Cathodic masstransport control Note that ba = βa/ln 10, and bc = βc/ln 10. Also, can be evaluated at or evaluated at or ; , and can be evaluated at or . Source: Ref 58

;

can be

References cited in this section 34. S. Barnartt, Electrochim. Acta, Vol 15, 1970, p 1313 35. K.B. Oldham and F. Mansfeld, Corros. Sci., Vol 13, 1973, p 813 53. J. Devay and L. Meszaros, Acta Chim. Acad. Sci. Hung., Vol 100, 1979, p 183 54. J.S. Gill, M. Callow, and J.D. Scantlebury, Corrosion, Vol 39, 1983, p 61 55. G.P. Rao and A.K. Mishra, J. Electroanal. Chem., Vol 77, 1977, p 121 56. L. Meszaros and J. Devay, Acta Chim. Acad. Sci. Hung., Vol 105, 1980, p 1 57. M.C.H. McKubre and B.C. Syrett, Corrosion Monitoring in Industrial Plants Using Nondestructive Methods, STP 908, G.C. Moran and P. Labine, Ed., American Society for Testing and Materials, 1986, p 433 58. R.W. Bosch, J. Hubrecht, W.F. Bogaerts, and B.C. Syrett, Corrosion, Vol 57, 2001, p 60–70


Electrochemical Noise Resistance Electrochemical noise analysis can provide a parameter called the electrochemical noise resistance, Rn (Ref 59, 60, 61, 62, 63, 64, 65). This parameter should be used in a manner analogous to polarization resistance. One electrode configuration that enables such a measurement involves connecting a zero-resistance ammeter (ZRA) between two nominally identical corroding electrodes immersed in the same solution. A third, nominally identical electrode can be immersed in solution and connected to the first two using a high-impedance voltmeter. This electrode serves as a “noisy” pseudoreference electrode. This approach is attractive in field applications due to the more rugged nature of the metallic electrode compared to laboratory electrodes but complicates the analysis because two uncorrelated potential sources (i.e., from the couple and the

pseudoreference) are measured in the collection of potential noise, Vn. Since Vn(meas) = (

+

)1/2, Vn(meas) must be divided by (Ref 66), to yield Vn(couple). Another alternative is a fourelectrode arrangement where the first pair is coupled through a zero-resistance ammeter to monitor current and the second pair is connected with a high-impedance voltmeter to sample an uncorrelated Vn(couple). Alternatively, a less noisy, conventional reference electrode may be utilized in the three-electrode arrangement. In this case, Vn(meas) and In(meas) are correlated. The reference electrode noise can be separately defined as the electrochemical voltage noise between two nominally identical reference electrodes (Ref 67). If the reference electrode noise is low, then the correction factor is not needed. In either case, the third electrode (reference electrode) is connected to the first two via a high-impedance voltmeter. These arrangements enable simultaneous recording of the galvanic current with time and the galvanic couple potential versus time. The standard deviation of the voltage noise divided by the standard deviation of the current noise has been proposed to yield a statistical parameter called the noise resistance Rn (Ref 59, 60, 61, 62, 63, 64, 65, 66, 68). Analysis of simulated noise data has led to the conclusion that the ratio of the standard deviations of the current and voltage noises measured between two identical electrodes can be normalized by surface area by multiplying by

(Ref 65): (Eq 27)

where σV(meas) is the standard deviation of the voltage noise, σI(meas) is the standard deviation of the current noise, and Av and AI are the surface areas of the corroding electrodes used for voltage and current measurement, respectively. Correlations between this parameter and conventionally determined polarization resistance as well as mass-loss-based corrosion rates have been obtained (Ref 60, 66). Unfortunately, experimental confirmation of the area normalization factor has not been extensively performed. Recall that in the case of a polarization resistance determined from E-iapp data or EIS data at the zero-frequency limit, measured resistance can be multiplied by electrode area and will yield the same area-normalized polarization resistance over a broad range of electrode areas. Moreover, the correlation has lacked a rigorous fundamental foundation for correlating Rn with corrosion rate despite the intuitive connection between σV and σI given by the proportionality factor Rn. The surface of one freely corroding electrode could be divided into area patches that experience fluctuations in interfacial resistance that produce changes in anodic and cathodic half-cell reaction rates in any one. The electrode potential must then change in each patch to drive the half-cell reactions such that the sum of all the anodic halfcell currents from all patches equals the sum of all cathodic half-cell currents, regardless of whether the source of cathodic half-cell current is from capacitive discharge or electrochemical reaction (Ref 64). Some global change in potential also occurs on the electrode. If the first electrode is now connected to a second electrode whose interfacial properties and global electrode potential do not change on their own at the same instant in time and by the same degree as on the first electrode, then a galvanic cell is momentarily created that induces a further difference in anodic and cathodic half-cell currents on the first electrode. Current now flows between the first and second electrode such that the sum of anodic and cathodic half-cell currents over all patches on both electrodes is equal. When the interfacial resistances return to normal values over all patches, the potential difference between the two electrodes is eliminated and so is the measurable current between the two electrodes. Bertocci argued that the external current fluctuation measured between two identical electrodes is identical to the fluctuation in one electrode (Ref 64, 65). Others have argued using concepts of mixed-potential theory that, at worst, the current sampled is only one-half of the total for equal-size electrodes. Theoretical relationships establishing the connection between Rn and Rp have been sought by several researchers (Ref 65, 69, 70, 71, 72), but their validity has been questioned. A great concern has been that the largest current peaks would occur during the most rapid voltage fluctuations since the electrode interface contains a capacitance through which current can be shorted (Ref 64, 65). Thus, when voltage fluctuations are rapid, the measured noise current will be shorted through the interfacial capacitance assuming a simple electrical equivalent circuit model consisting of two parallel resistor-capacitor networks describing the interface for each electrode connected in series through Rs. This situation would lead to the lowest impedance between the two electrodes during the most rapid voltage fluctuations that, in turn, produce the greatest current fluctuations. The theoretical maximum measured current would be given by the voltage fluctuation divided by

Rs. The outcome would be a statistical noise-resistance parameter that is proportional to, or heavily influenced by, higher- frequency data. Indeed, Rn was found to equal an absolute impedance at some frequency that depended on the frequency of the voltage fluctuations and the RC time constant of the electrode interface in one study of simulated noise (Ref 64). Unfortunately, a Rn value obtained at high frequency would be smaller in magnitude than the Rp obtained at the zero-frequency limit. Hence, it would not represent the desired zerofrequency limit interfacial resistance, Rp. Indeed, such underestimations in the true value of Rp have been observed experimentally (Ref 66, 68). Recently, a more rigorous theoretical and experimental analysis has been made comparing the spectral noise resistance obtained at each frequency with both the polarization resistance obtained from the zero-frequency limit of impedance magnitude data, |Z(ω = 0)|, as well as the frequency-dependent impedance of two electrodes (Ref 73, 74, 75, 76, 77). The spectral noise resistance Rsn(ω) was determined by taking the square root of power spectral density of the voltage noise (V2/Hz)1/2 and dividing it by the square root of power spectral density of the current noise (A2/ Hz)1/2 at each frequency using the same two- electrode arrangement as discussed previously (Ref 76, 77): (Eq 28) Rsn(ω) is proportional to the magnitude of the cell impedance, |Z(ω)| in the two-electrode arrangement (Ref 76, 77). The proportionality factor is unity in the case of identically sized electrodes in a two-electrode cell with identical impedances and a noiseless reference electrode (Ref 76, 77). Therefore, the spectral noise resistance at the zero-frequency limit could equal the interfacial impedance at the zero-frequency limit |Z(ω = 0)| in the theoretical case of identical electrode impedances with negligible Rs. Figure 6 gives data for identical iron electrodes in 1 M Na2SO4 with an iron reference electrode. Here Rsn(ω) = |Z(ω)| due to the noisy reference electrode. Thus 2|Z(ω)| and Rsn(ω) appear to be similar. It is well known that in many instances |Z(ω = 0)| equals Rp. Even Rn may equal Rsn(ω = 0) = |Z(ω = 0)| = Rp if |Z(ω)| equals Rp in the frequency regime dominating the Rn value. The frequency range dominating the Rn value is determined by several factors, but this statement is more likely to be true if |Z(ω)| and Rsn(ω) both exhibit long low-frequency plateaus over a broad frequency range that encompasses the fmin and fs utilized in the Rn measurement. Here fmin is given by the total sampling time, T, where fmin = 1/T and fs equals the data sampling rate. Rn typically varies with fs and underestimates |Z(ω = 0)|. Unfortunately, Rsn(ω → 0) does not equal Rp in the zero-frequency limit under many other conditions, such as when log (Rs/Rp) > 0 or in the case of very noisy reference electrodes (Ref 76, 77). Moreover, Rsn(ω) can be dominated by the properties of the high-impedance electrode in the case of dissimilar electrode impedances that are equally noisy, but this is not always the case. For instance, the low-impedance electrode in a two- electrode cell with a third reference electrode can be sensed by Rsn(ω) if the higher impedance electrode is much noisier than the low-impedance electrode (Ref 76, 77). Recent attempts have been made to address circumstances where Rsn(ω) lies in between |Z(ω)|1 and |Z(ω)|2 representing the high- and low-impedance electrodes. Methods have been suggested for sensing the current fluctuations on both electrodes (Ref 72). The reader is referred to these articles for more information.

Fig. 6 Rsn(ω) versus frequency compared to two times the impedance |Z(ω)| versus frequency for iron in 1 M Na2SO4 at pH 4 with a “noisy” iron reference electrode. Impedance measurements performed in a two-electrode cell with two iron electrodes produced 2|Z(ω)|. Rsn(ω) calculated to equal |Z(ω)| for the case of two iron electrodes coupled through a zero-resistance ammeter and a third iron electrode as reference electrode. Source: Ref 77

References cited in this section 59. D.A. Eden, A.N. Rothwell, and J.L. Dawson, Paper 444, Proc. Corrosion Conference, 1991, National Association of Corrosion Engineers, 1991 60. J.L. Dawson, Electrochemical Noise Measurement for Corrosion Applications, STP 1277, J. Kearns, J.R. Scully, P.R. Roberge, D.L. Reichert, and J.L. Dawson, Ed., American Society of Testing and Materials, 1996, p 3–35 61. D.A. Eden, K. Hladky, D.G. John, and J.L. Dawson, Paper 276, Proc. Corrosion Conference, 1986, National Association of Corrosion Engineers, 1986 62. D.A. Eden and A.N. Rothwell, Paper 292, Proc. Corrosion Conference, 1992, National Association of Corrosion Engineers, 1992 63. A.N. Rothwell and D.A. Eden, Paper 223, Proc. Corrosion Conference, 1992, National Association of Corrosion Engineers, 1992 64. U. Bertocci, Electrochemical Noise Measurement for Corrosion Applications, STP 1277, J. Kearns, J.R. Scully, P.R. Roberge, D.L. Reichert, and J.L. Dawson, Ed., American Society of Testing and Materials, 1996, p 39–58 65. U. Bertocci and F. Huet, Corrosion, Vol 51, 1995, p 131 66. D.L. Reichert, Electrochemical Noise Measurement for Corrosion Applications, STP 1277, J. Kearns, J.R. Scully, P.R. Roberge, D.L. Reichert, and J.L. Dawson, Ed., American Society of Testing and Materials, 1996, p 79–89 67. P.C. Searson and J.L. Dawson, J. Electrochem. Soc., Vol 135 (No. 8), 1988, p 1908 68. F. Mansfeld and H. Xiao, Electrochemical Noise Measurement for Corrosion Applications, STP 1277, J. Kearns, J.R. Scully, P.R. Roberge, D.L. Reichert, and J.L. Dawson, Ed., American Society of Testing and Materials, 1996, p 59–78 69. G.P. Bierwagen, J. Electrochem. Soc., Vol 141, 1994, p L155 70. F. Mansfeld and H. Xiao, J. Electrochem. Soc., Vol 141, 1994, p 1403 71. F. Huet, J. Electrochem. Soc., Vol 142, 1995, p 2861 72. J.F. Chen and W.F. Bogaerts, Corros. Sci., Vol 37, 1995, p 1839 73. H. Xiao and F. Mansfeld, J. Electrochem. Soc., Vol 141, 1994, p 2332 74. F. Mansfeld and H. Xiao, J. Electrochem. Soc., Vol 140, 1993, p 2205 75. F. Mansfeld, L.T. Han, and C.C. Lee, J. Electrochem. Soc., Vol 143, 1996, p L286

76. U. Bertocci, C. Gabrielli, F. Huet, and M. Keddam, J. Electrochem. Soc., Vol 144, 1997, p 31 77. U. Bertocci, C. Gabrielli, F. Huet, and M. Keddam, J. Electrochem. Soc., Vol 144, 1997, p 37


Electrochemical Methods for the Study of Galvanic Corrosion Rates Methods Based on Mixed-Potential Theory. The thermodynamic tendency for galvanic corrosion may be determined from the electromotive series (Ref 4, 5) or from the construction of a galvanic series (Ref 79), as discussed in ASTM G 82, “Guide for Development and Use of a Galvanic Series for Predicting Galvanic Corrosion Performance.” Galvanic corrosion rates can also be determined from mixed-potential theory as shown in Fig. 7. In the case of bimetal or multimetal galvanic attack in which two or more metals are electrically in contact with one another, there is in theory a minimum of two cathodic and two anodic reactions. One of each of these reactions is occurring on each metal. In this case, the more noble of the two metals is cathodically polarized, and its anodic reaction rate will thus be suppressed. Conversely, the less noble or anodic material is anodically polarized, and the anodic reaction rate is accelerated. The mixed potential (the galvanic couple potential, Ecouple) of the galvanic couple and the resulting galvanic current can be uniquely determined from the sums of all of the individual anodic and cathodic currents obtained for each material at each potential when: ΣiaAa = ΣicAc at Ecouple

(Eq 29)

where ia and ic are the anodic and cathodic current densities (μA/cm2), respectively, and Aa and Ac are the anodic and cathodic areas (cm2).

Fig. 7 Potential-current relationships for the case of a galvanic couple between two corroding metals. Iron is the more noble metal; zinc is less noble metal. The galvanic couple potential is defined by the potential where the sum of the anodic currents equals the sum of the cathodic currents for all of the reactions on all of the metals in the couple. The galvanic couple potential can be determined either by direct measurement with a reference electrode or from polarization data if: (1) polarization data are available for each material in the galvanic couple, (2) the area of each metal is known, and (3) the current distribution is uniform. Once Ecouple is determined, the galvanic corrosion rate may be estimated for the metal of interest in the couple if a relationship such as given in Eq 7 is known for that metal. In simple bimetal cases, direct superposition of polarization data (corrected for wetted surface area) yields Ecouple and the galvanic corrosion rate (Ref 79). However, because applied currents instead of true anodic or cathodic currents are measured in any polarization experiment, the E-log(I) superposition technique will introduce the least error when the cathodic reduction reaction rate on the anode is negligible and the anodic oxidation reaction rate on the cathode is negligible at the galvanic couple potential. Obviously, when the open circuit potential OCPs of the anode and cathode are similar, error is more likely. Fortunately, galvanic corrosion may be less significant in these cases. In addition, special care must be taken in the procedures used to develop the polarization data (Ref 79), especially if time effects are to be taken into consideration when evaluating long-term galvanic corrosion behavior. Direct Measurement of Galvanic Corrosion Rates. A more straightforward procedure involves immersing the two dissimilar metals in an electrolyte and electrically connecting the materials together using a zero-resistance ammeter to measure the additional galvanic current (Ref 9, 80, 81). In this method, the galvanic current is directly determined as a function of time. The galvanic corrosion rate so determined is the additional corrosion

created with the couple and will not equal the true corrosion rate. This is given by the sum of the galvanic corrosion rate and the corrosion rate under freely corroding conditions unless the latter is negligible. Recall that the corrosion rate of the uncoupled anode is undetermined by this method since an equal cathodic reaction rate is occurring on the same surface. A reference electrode connected to the galvanic couple can be used to determine the galvanic couple potential. ASTM standards do not exist for direct, mixed-potential theory, or scanning potential probe methods. Potential probe methods may be used to determine and map the local ionic currents associated with galvanic corrosion cells between dissimilar metals or heterogeneities on complex alloy surfaces (Ref 82, 83, 84, 85). In the most straightforward application, the local potential is mapped over a planar electrode oriented in the x-y plane to give an indication of local current. The basic concept is that the ionic current density in three dimensions can be mapped by either scanning an array of reference electrodes or by vibrating a single electrode. The orthogonal ionic current flow can be expressed in terms of solution conductivity and the gradient in potential in the solution above the galvanic couple: i = -κ(

E)

(Eq 30)

where κ is the solution conductivity and E = δ1dE/dx + δ2dE/dy + δ3dE/dz where x, y, and z define axes in a coordinate system and δ1, δ2, and δ3 are unity vectors. The advantage of the vibrating technique is that minor differences between the reference potential of separate electrodes is eliminated by using a single vibrating reference electrode. The ionic current density so recorded is the component of current density flowing perpendicular to isopotential lines in solution established due to the galvanic couple and the established potential gradient. Therefore, a map of local current can be constructed by scanning over a planar electrode in the x-y plane, where z is the vertical distance in the solution above the electrode. Locations of high local current imply significant galvanic interactions.

References cited in this section 4. D.A. Jones, Principles and Prevention of Corrosion, 2nd ed., Macmillan Publishing, 1996 5. M.G. Fontana and N.D. Greene, Corrosion Engineering, McGraw-Hill, 1978 9. R. Baboian, Ed., Electrochemical Techniques for Corrosion, National Association of Corrosion Engineers, 1977 79. H. Hack and J.R. Scully, Corrosion, Vol 42 (No. 2), 1986, p 79 80. R. Baboian, Galvanic and Pitting Corrosion-Field and Laboratory Studies, STP 576, R. Baboian et al., Ed., American Society for Testing and Materials, 1974, p 5 81. F. Mansfeld and J.V. Kenkel, Galvanic and Pitting Corrosion-Field and Laboratory Studies, STP 576, R. Baboian et al., Ed., American Society for Testing and Materials, 1974, p 20 82. R.G. Kasper and C.R. Crowe, Comparison of Localized Ionic Currents as Measured from 1-D and 3-D Vibrating Probes, Galvanic Corrosion, STP 978, H.P. Hack, Ed., American Society for Testing and Materials, 1988, p 118 83. R.G. Kasper and C.R. Crowe, Comparison of Localized Ionic Currents as Measured from 1-D and 3-D Vibrating Probes, J. Electrochem. Soc., Vol 33, 1986, p 879 84. H.S. Isaacs, The Measurement of Galvanic Corrosion of Soldered Copper Using the Scanning Vibrating Electrode Technique, Corros. Sci., Vol 28, 1988, p 547 85. V.S. Voruganti, H.B. Huft, D. Degeer, and S.A. Bradford, Scanning Reference Electrode Technique for the Investigation of Preferential Corrosion of Weldments in Offshore Applications, Corrosion, Vol 47, 1991, p 343


Electrochemical Methods for the Definition of Conditions Where Corrosion Rates Are High Pitting and crevice corrosion are associated with the local breakdown of passivity, often in halide-containing solutions (Ref 86). Passive corrosion rates are often very low and are not discussed further. Localized corrosion occurs at rates an order of magnitude greater than the rates during passive dissolution. Therefore, it is often of great interest to determine the conditions under which localized corrosion occur instead of the exact propagation rates, which are usually intolerable. Electrochemical tests for evaluating the susceptibility of a material to pitting and to crevice corrosion include potentiodynamic, potentiostatic, scratch potentiostatic, potentiostaircase, tribo-ellipsometric methods, pit-propagation rate curves, galvanostatic, and electrochemical noise measurements (Ref 87, 88). These methods identify critical potential, temperature, and environmental conditions associated with the thresholds between low rates (e.g., passive dissolution at 1 μA/cm2 or less than 1 mpy depending on alloy) and high rates of corrosion (e.g., locally as high as 1 A/cm2). Two critical potentials of significant importance are the breakdown (Ebd) and protection (Eprot) or repassivation potential. The former is the potential at which pits or crevices develop or become stabilized. The latter is often defined as the least noble potentials at which pits, once formed, will propagate. Below this potential, pits repassivate. A simple potentialbased criterion for local corrosion is the comparison of Ecorr with such a threshold potential associated with breakdown of passive films. A useful concept for safety against the transition to local corrosion is that Ecorr « Ebd and Eprot where Ebd is often associated with stabilization of local corrosion and Eprot is associated with repassivation. When this situation exists, the chance of local high-rate corrosion processes is often low. Therefore, it is important to determine these two potentials. Cyclic Potentiodynamic Polarization Methods to Determine Ebd and Eprot. ASTM G 61 (Ref 89) describes a procedure for conducting cyclic potentiodynamic polarization measurements to determine relative susceptibility to localized corrosion. The method is designed for use with iron- or nickel-base alloys in chloride environments. In this test, a cyclic anodic polarization scan is performed at a fixed voltage scan rate. Figure 8(a) shows the cyclic potentiodynamic method. Particular attention is focused on two features of cyclic polarization behavior (Fig. 8a). The first is the potential at which the anodic current increases significantly with applied potential (the breakdown potential, Ebd). In general, the more noble this potential, obtained at a fixed scan rate in this test, the less susceptible the alloy is to the initiation of localized attack. The second feature of great interest is the potential at which the hysteresis loop is completed during reverse polarization scan after localized corrosion propagation. This potential is often taken as the repassivation potential or protection potential, Eprot. In general, once initiated, localized corrosion sites can propagate only at some potential more positive than the potential at which the hysteresis loop is completed (when determined at a fixed scan rate). In other words, repassivation will occur at more negative potentials even after localized corrosion initiation. Therefore, the more positive the potential at which the hysteresis loop is completed, the less likely that localized corrosion will propagate. ASTM G 61 (Ref 89) discusses cyclic polarization behavior for Hastelloy C276 and 304 stainless steel in 3.5% sodium chloride solution. Based on this criterion, it is evident that Hastelloy C-276 is more resistant to localized corrosion in this environment than AISI 304 stainless steel.

Fig. 8 (a) Cyclic potentiodynamic polarization curve. (b) Galvanostatic potential-time curve for a material. (c) Potentiostatic current-time curve for a previously passivated surface which pits at E1 < EBD < E2. (d) Potentiostatic current-time curve for active surface. The protection potential is found as E3 < Eprot < E2. Source: Ref 86 Complications with Cyclic Potentiodynamic Polarization Methods. Although the cyclic method is a reasonable method for screening variations in alloy composition and environments for pitting susceptibility, the cyclic potentiodynamic polarization method has been found to have a number of shortcomings (Ref 87, 88, 89, 90, 91, 92). One major problem concerns the effect of the potential scan rate. The values of both Ebd and Eprot are a strong function of the manner in which the tests are performed, particularly the potential scan rate employed. Experimental values of Ebd are linked to the induction time required for pitting. Another complication arises from allowing too much pitting propagation to occur before reversing the scan direction. This either alters the localized chemistry in pits, affects pit depth, or both. Pit depth alters the diffusion length associated with the dilution of pit chemistry necessary for repassivation. This factor affects the polarization behavior after the reversal in the direction of scanning and influences Eprot. From an engineering standpoint, metal surfaces held at potentials below the repassivation potential of the deepest pits should be safe against stable pit propagation.

That is to say, stable pits should not propagate. It has been found that Ebd observed after potentiostatic holds or the slowest scan rates approaches the protection potential found after minimal pit growth (Ref 92). This suggests the existence of a single critical potential for pit growth in the absence of crevices or other occluded sites (Ref 92). Potentiostatic and Galvanostatic Methods for Localized Corrosion. The shortcomings of the cyclic potentiodynamic polarization method have become the basis for several other electrochemical techniques. Other methods are shown in Fig. 8. In the galvanostatic or galvanostaircase technique (Fig. 8b), potential is measured versus time at various constant applied currents that are incrementally increased in steps, then reversed, and decreased. In the case of passive materials, a potential rise during galvanostatic testing indicates passive film growth, while a decline indicates breakdown and growth of local corrosion sites. In the galvanostaircase technique, the current is step increased. Potential measurements are made until the time rate of change in potential approaches 0. These forward and reverse potential-current density data are extrapolated to the zero current density to obtain Ebd and Eprot. The technique is described by ASTM G 100 “Standard Test Method for Conducting Cyclic Galvanostaircase Polarization Tests” (Ref 93) as a test method for aluminum alloys. Potentiostatic methods can overcome the inherent problems involving scan rate. A more conservative estimate of Ebd can be obtained by polarizing individual samples for long periods of time at potentials above and below the values of Eprot and Ebd previously determined from the potentiodynamic method (Fig. 8c). Eventual initiation is indicated by a current increase. In another approach (Fig. 8d), initiation of pits is intentionally induced by applying a “stimulation” potential well above Ebd and then quickly shifting to a preselected potential below that value. If this second applied potential is above Eprot, propagation of the existing pits will continue and the current will increase. However, at potentials less than Eprot the pits will eventually repassivate and the current will subsequently decrease with time. The critical potential for pitting is defined as the most noble potential at which pits repassivate after the stimulation step. This approach is covered in ASTM F 746 “Test Method for Pitting or Crevice Corrosion of Metallic Surgical Implant Materials” (Ref 94). Determination of Eprot by Potential Step- Down or Scan-Down Methods. As stated previously, Ebd and Eprot often depend strongly on the method by which they are determined and, therefore, do not uniquely define intrinsic material properties. The Eprot values determined from the scanning method can be complicated by scan rate, pit size or depth, vertex potential/ current, polarization curve shape, and specimen geometry (Ref 95, 96). Some investigators have found more consistent Eprot values after a critical charge has passed, while others report a single critical potential (Ref 92). Often this potential is difficult to determine and has been taken at various points of the polarization curve (Ref 98). What is needed is a method for determination of Eprot that defines a conservative value of this potential that likely reflects a true pit or crevice repassivation potential. Tsujikawa has developed a method for determination of Eprot from previously grown pits and crevices (Ref 99, 100). This method and its variations have been successfully used by several research groups and associates the critical potential for repassivation Eprot with the need to grow local corrosion sites to a critical minimum size (Ref 101). The method is an enhancement of the determination of Eprot from potentiodynamic E-I scans that involve scan reversal to the point where pits are repassivated. In the potential step-down method, the potential is first set at a high enough value to induce and grow stable pits to the specified size. The potential can then be stepped down or scanned downward while the pit propagation current is recorded. Subsequently, the potential may be held after initiating a pit in order to determine the time until repassivation (Ref 102). Moreover, long potential holds at selected applied potentials enables confirmation of a true repassivation potential often indicated by an abrupt decrease in current density at the time when the chemical conditions favoring pit stability (often expressed as some fraction of the saltsaturation concentration) are no longer maintained. Abrupt transitions in finite repassivation time toward infinity with increasing hold potential may indicate that conditions favorable to pit growth are sustained. Another benefit of determining Eprot from controlled growth of pits and crevices is that the repassivation potential can be determined from pits of preselected sizes that can be controlled by the duration of the potential holds. The method has been successfully applied to nickel-base alloys, stainless steels, and aluminum-base alloys. This method is being standardized by the committee for Japanese Industrial Standards. When a onedimensional pit or pencil electrode is tested, the pit propagation kinetics can be recorded from a single pit (Ref 103). Whether pit growth is ohmic, mass transport, or charge-transfer controlled may also be determined. Moreover, the effects of various material and solution parameters (e.g., flow rate, conductivity, and solution composition) on pit propagation can be determined.

The Scratch-Repassivation Method for Localized Corrosion. One additional potentiostatic technique to be mentioned in the area of localized corrosion involves the scratch method (Ref 96). In this method, the alloy surface is scratched at a constant potential and the current is measured as a function of time. The potential dependencies of the induction time and the repassivation time are determined by monitoring the current change over a range of different potentials. This is shown in Fig. 9. From this information the critical pitting potential, thought to be less than Ebd determined by potentiodynamic scan, can be found. Other methods of studying localized corrosion are also available (Ref 86).

Fig. 9 Potential versus time plot of scratch test illustrating a possible location of the critical potential, Ec, as it relates to the induction time and the repassivation time. Source: Ref 86 Statistical Distributions in Critical Potentials. Ebd is typically observed to be a significantly “statistically distributed” property compared to the repassivation potential. This is often observed when critical potentials are determined by potentiodynamic scanning, but can be observed during potentiostatic tests where the pit incubation time and subsequent survival probabilities are also seen to be distributed. Eprot distributions are often attributed to distributions in pit size. However, Eprot is not distributed when pits are uniformly large (Ref 97)

and may not depend on the depth of corrosion damage at all during crevice corrosion when a crevice former controls mass transport (Ref 104). Statistical distributions in pitting potentials have been observed for AISI 304 (Ref 105), Fe-Cr-X (where X = niobium, molybdenum, or titanium) alloys (Ref 106), AISI 316 (Ref 107), titanium (Ref 108), high-purity aluminum (Ref 109), 2024-T3 (Ref 110), and amorphous aluminum-base alloys (Ref 111). This means that significant variations are seen when a number of specimens are tested under identical conditions. These distributions have been attributed to the stochastic nature of pitting (Ref 107), distributions in oxide-film cation vacancy transport properties (Ref 112), the effects of potential on the nature of the eligible sites for the metastable pit nucleation process (Ref 105), and the population of fatal flaws or pit initiating defects on an electrode surface (Ref 113). The possibility that oxide defects at densities exceeding millions of sites/cm3 produces specimen-to-specimen variations appears unlikely for specimens with surface areas of the order of a few cm2. It appears to be more reasonable to expect that sample-to-sample variations are, instead, associated with distributions in the population of micrometer-scale defects such as sulfide inclusions in stainless steel (Ref 114) and constituent particles in aluminum alloys (Ref 112) that control pit nucleation intensity and produce micropit-to-micropit variations in growth rates and transport characteristics (e.g., Deff, pit shape) for a large population of metastable pits grown at a single potential. A statistical distribution of peak Ipit/rpit ratios (where Ipit is the peak pit current and rpit is pit radius) was observed for a large population of pits at a fixed potential in high-purity aluminum (Ref 109). Although not rigorously proven, it is reasonable to argue that a slightly lower Deff or slightly larger Ipit/rpit ratio for a given pit might lead to pit stabilization at differing applied potentials on an electrode with its own unique population of metastable pits. Moreover, such behavior is likely controlled by micrometer- scale defects spaced at tens of micrometer lateral separation distances across planar electrode surfaces. Investigators who seek information on critical potentials for engineering use should consider appropriate specimen sizes relative to critical defect densities and spacings. There are no ASTM standards that address these issues. However, the size of test specimens recommended in many ASTM standards is conservatively large, leading to conservative values of Ebd. Electrochemical noise (EN) methods are used increasingly as a nondestructive tool for evaluating susceptibility to localized corrosion as well as stress-corrosion cracking (SCC), particularly in field or process plant applications (Ref 115, 116, 117, 118, 119, 120). Electrochemical noise methods are appealing because they may be conducted at open circuit without perturbing the corroding system. However, no consensus currently exists as to the most appropriate test procedure or analysis method, or how well results correlate with coupon exposures. The transient development of bare metal at newly formed pit or cracking sites as a result of temporary propagation and repassivation can result in potential noise (open-circuit EN), current noise (potentiostatic EN), or both. In the latter case, the current noise is measured with a zero-resistance ammeter (ZRA) used to monitor a galvanic couple consisting of two identical electrodes while the potential noise comes from the reference electrode (or third metallic electrode) monitoring the couple potential. The noise signal, hereafter referred to as a potential or current time record, is caused by the galvanic couple formed between the very small anode sites corroding at current densities approaching 10 A/cm2 in pitting corrosion and the much larger remaining cathode surface operating at lower cathodic current densities (e.g., 10 μA/ cm2). Regarding pitting phenomena, a negative shift in measured potential is observed (OCP and galvanic couple EN), an increase in current is observed for potentiostatic EN, and current fluctuations of either polarity are observed for galvanic couple EN. Several analysis methods exist including electrochemical (Ref 121), statistical (Ref 119, 120), spectral (Ref 115, 116, 117, 118, 119, 120, 121), and autocorrelation (Ref 117). Some of the approaches for determination of spectral noise resistance during uniform corrosion are discussed in the section on polarization resistance. Electrochemical analysis may also enable the determination of pit sizes from the charge associated with each pitting event (Ref 109) or from attempting to determine pit current densities. Statistical analyses include determining the root mean square (rms), variance, and standard deviations of the EN voltage or current time records, as well as a noise resistance, REN, taken as the ratio of the rms or standard deviation of the potential and current time records acquired over various periods of time (Ref 105). Spectral analysis consists of Fourier transformation of EN data acquired in the time domain to create a power spectral density plot (Ref 117, 118, 120, 122). Qualitative assessment of the benefits of inhibitors or the effects of process stream variations are made possible by comparing statistical or spectral results. Quantitative analysis and predictive capability still require further development. Use of Shot Noise Methods During Pitting. In the prepitting stage, multiple subcritical pitting events are often observed. The current time record shows exponential decaying or rising transients associated with discrete

pitting events. In many systems, these prepitting events occur at potentials far more negative than Ebd and sometimes Eprot. Numerous investigators have proposed that the ratio of peak pit current to pit radius during metastable pitting or product of pit current density times pit radius (Ipit/rpit or ipit · rpit) provides an insightful parameter that can be used to characterize the risk of stable pit propagation based on the critical acidified chemistry theory of Galvele (Ref 123). When anodic current is monitored near the pitting potential, multiple overlapping metastable pitting events are often detected as current spikes at potentials just below Ebd. Electrochemical noise analysis provides one method for extracting information on pitting from such complicated current time records. That is, the power spectrum for a population of spikes of similar amplitude and duration is the sum of the spectra for each spike and should have the same shape as an individual spike. It has been proposed to use shot noise analysis to calculate the pit charge from power spectral density (PSD) obtained from a current-time record (Ref 124, 125, 126). According to this approach, pit charge, qpit, and through Faraday's Law the pit radius, rpit, for the pit events that dominate the current time record, can be determined from the PSD of a pitting current noise spectrum. The PSD of current fluctuations for the case of randomly occurring, exponentially decaying current spikes associated with metastable pitting events during the prepitting stage can be described by (Ref 123): (Eq 31) where f is the spectrum frequency, Δf is the frequency resolution given by the data acquisition frequency divided by the number of data points collected, qpit is the pit charge, Imean is the mean current from a I-t record containing multiple current fluctuations associated with pit events after subtraction of any background passive current, I is the amplitude of pit current fluctuations, and τ is the time constant for individual exponentially decaying current transients that comprise the current fluctuations. Hence, the low-frequency limit of a log I2/Δf versus log f plot provides information on qpit for known Imean. The frequency of the intercept (or breakpoint frequency) of the sloping part of the high- frequency PSD with the low-frequency plateau, fbpt, provides information on the time constant for exponentially decaying pitting events, τ; fbpt = πτ. Therefore, qpit and τpit may be extracted from current PSD plots constructed from current-time data containing many overlapping metastable pitting events. This has been successfully done for aluminum alloys (Ref 127, 128). Unfortunately, the approach has its shortcomings. The parameters recovered are heavily biased toward smaller, more numerous pitting events that dominate the current time record. The biggest and/or fastest growing pits that might readily form stable propagating pits may not be detected. However, the method does enable determination of pit sizes at various potentials and allows the comparison of one environment, temperature, or set of plant conditions to another. No ASTM standard currently exists on this topic.

References cited in this section 86. Z. Szklarska-Smialowska, Pitting Corrosion of Metals, National Association of Corrosion Engineers, 1986 87. B.E. Wilde, Corrosion, Vol 28, 1972, p 283 88. B.C. Syrett, Corrosion, Vol 33, 1977, p 221 89. “Test Method for Conducting Cyclic Potentiodynamic Polarization Measurements for Localized Corrosion Susceptibility of Iron-, Nickel-, of Cobalt-Based Alloys,” G 61, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 90. N. Pessall and C. Liu, Electrochim. Acta, Vol 16, 1971, p 1987 91. B.E. Wilde and E. Williams, J. Electrochem. Soc., Vol 118, 1971, p 1057 92. N.G. Thompson and B.C. Syrett, Corrosion, Vol 48 (No. 8), 1992, p 649

93. “Standard Method for Conducting Cyclic Galvanostaircase Polarization,” G 100, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 94. “Test Method for Pitting or Crevice Corrosion of Metallic Surgical Implant Materials,” F 746, Annual Book of ASTM Standards, ASTM 95. B.E. Wilde and E. Williams, Electrochem. Acta, Vol 16, 1971, p 1671 96. K.K. Starr, E.D. Verink, Jr., and M. Pourbaix, Corrosion, Vol 2, 1976, p 47 97. N. Sridhar and G.A. Cragnolino, Corrosion, Vol 49, 1993, p 885 98. M. Yasuda, F. Weinburg, and D. Tromans, J. Electrochem. Soc., Vol 137, 1990, p 3708 99. S. Tsujikawa, T. Hasamatsu, and G. Boshoku, Corros. Eng., Vol 29, 1980, p 37 100.

T. Shinohara, S. Tsujikawa, and G. Boshoku, Corros. Eng., Vol 39, 1990, p 238

101.

S. Tsujikawa, Z. Heng, Y. Hisamatsu, and G. Boshoku, Corros. Eng., Vol 32, 1983, p 149

102. D. Dunn and N. Sridhar, Critical Factors in Localized Corrosion II, Proc. Electrochemical Society 95-15, P.M. Natishan, R.G. Kelly, G.S. Frankel, R.C. Newman, Ed., 1996, p 79–90 103. G.T. Gaudet, W.T. Mo, T.A. Hatton, J.W. Tester, J. Tilly, H.S. Isaacs, and R.C. Newman, Am. Inst., Chem. Eng. J., Vol 32 (No. 60), 1986, p 949 104.

B.A. Kehler, G.O. Ilevbare, and J.R. Scully, Corrosion, Vol 57 (No. 12), 2001, p 1042–1065

105.

P.C. Pistorius and G.T. Burstein, Philos. Trans. R. Soc. London, Vol 341, 1992, p 531

106.

T. Shibata, Trans. ISIJ, Vol 23, 1983, p 785

107.

T. Shibata and T. Takeyama, Corrosion, Vol 33, 1977, p 243

108.

T. Shibata and Y.-C. Zhu, Corros. Sci., Vol 36 (No. 1), 1994, p 153–163

109.

S.T. Pride, J.R. Scully, and J.L. Hudson, J. Electrochem. Soc., Vol 141, 1994, p 3028

110.

G.O. Ilevbare, J.R. Scully, J. Yuan, and R.G. Kelly, Corrosion, Vol 56 (No. 3), 2000, p 227–242

111.

J.E. Sweitzer, G.J. Shiflet, and J.R. Scully, Tri-Service Military Corrosion Conf. Proc., 2002

112. D.D. Macdonald, Critical Factors in Localized Corrosion, The Electrochemical Society Proceedings Series, PV 92-9, G.S.Frankel and R.C. Newman, Ed., 1992, p 144 113.

T. Suter and R.C. Alkire, J. Electrochem. Soc., Vol 148 (No. 1), 2001, p B36–B42

114.

G.T. Burstein and G.O. Ilevbare, Corros. Sci., Vol 38, 1996, p 2257–2265

115.

K. Hladky and J.L. Dawson, Corros. Sci., Vol 22, 1982, p 231–237

116.


117. H.S. Bertocci, Proc. Second International Conf. Localized Corrosion, NACE-9, H. Isaacs, U. Bertocci, J. Kruger, and S. Smialowska, Ed., National Association of Corrosion Engineers, 1990 118.

H.S. Bertocci and J. and Kruger, Surf. Sci., Vol 101, 1980, p 608–618

119. D.A. Eden and A.N. Rothwell, “Electrochemical Noise Data: Analysis and Interpretation,” Paper 92, Proc. Corrosion Conference, 1992, National Association of Corrosion Engineers, 1992 120.

P.C. Searson and J.L. Dawson, J. Electrochem. Soc., Vol 135, 1988, p 1908–1915

121. D.E. Williams, J. Stewart, and B.H. Balkwill, Proc. Symposium on Critical Factors in Localized Corrosion, Vol 92-9, G.S. Frankel and R.C. Newman, Ed., Electrochemical Society, 1992, p 36 122. S.T. Pride, J.R. Scully, and J.L. Hudson, Electrochemical Noise Measurement for Corrosion Applications, STP 1277, J.R. Kearns and J.R. Scully, et al., Ed., American Society for Testing and Materials, 1996, p 307 123.

J.R. Galvele, J. Electrochem. Soc., Vol 123, 1976, p 464

124. S. Turgoose and R.A. Cottis, “Corrosion Testing Made Easy, Electrochemical Impedance and Noise” (Houston, TX), NACE International, 2000 125.

R.A. Cottis, Corrosion, Vol 57, 2001, p 265

126. C. Gabrielli, F. Huet, M. Keddam, and R. Oltra, Localized Corrosion as a Stochastic Process, NACE—9 Advances in Localized Corrosion, H. Isaacs, U. Bertocci, J. Kruger, and S. Smialowska, Ed., National Association of Corrosion Engineers, 1990, p 93–108 127. J.R. Scully, S.T. Pride, H.S. Scully, and J.L. Hudson, Critical Factors in Local Corrosion II, Proc. Electrochemical Society, Vol 95-15, R.G. Kelly et al., Ed., 1996, p 15 128.

R.A. Cottis and S. Turgoose, Mater. Sci. Forum, Vol 2, 1995, p 192–194


Evaluation of Intergranular Corrosion Rates Most methods of evaluating intergranular corrosion test for susceptibility as opposed to intergranular corrosion rate. There are many acid-exposure-type tests that rely on visual and metallographic examination to evaluate susceptibility. One example of an electrochemical test that examines susceptibility is described in this section. Stainless steel alloys that exhibit susceptibility can be expected to experience high rates of intergranular corrosion in certain service environments. Evaluation of Alloy Sensitization. Measurement of the coulombs generated during the electrochemical polarization of a material from the passive range to the active corrosion potential can be used to quantify the susceptibility to intergranular attack associated with the precipitation of chromium carbides and chromium

nitrides at grain boundaries (Ref 129, 130, 131). The method is described by ASTM G 108 (Ref 132). A modification of this procedure, called the double-loop electrokinetic repassivation test (Ref 133, 134), involves a potentiodynamic polarization of the metal surface initially from the open-circuit potential in the active region to a potential in the passive range. This is immediately followed by a reverse polarization in the opposite direction back to the open-circuit potential. The second method is less dependent on surface finish and precise knowledge of the grain size. Both variations of the method are shown in Fig. 10. In the latter method, the degree of sensitization is determined by the ratio, (Ir/Ia), of the maximum current generated in the reactivation, Ir, or reverse scan compared to that generated in the initial anodic scan, Ia. The procedure is contingent on the presence of anodic current during the reactivation scan that results mainly from incomplete passivation of the zone adjacent to the grain boundaries that is depleted of chromium due to carbide precipitation at grain boundaries. For nonsensitized material, the passive film remains essentially intact during the reverse scan and the magnitude of the reactivation polarization peak remains small. For the same reasons, the charge Q (obtained from integration of current versus potential for a known voltage scan rate) in the single-loop method is small for nonsensitized material. As a refinement to the method the charge is normalized by the grain- boundary area (GBA) because this is the area from which most of the current arises in the single reactivation scan (Ref 131): P = Q/GBA

(Eq 32) (Eq 33)

where P is the reaction charge density associated with the sensitized area (coulombs/cm2), As is the wetted specimen surface area, GBA is the grain-boundary area, and GS is the ASTM grain size in accordance with ASTM E 112, “Test Methods for Determining the Average Grain Size.”

Fig. 10 Two procedures for anodic reactivation polarization testing. (a) Clarke method. (b) Akashi method. Source: Ref 133, 134 The same procedure can be used to normalize the ratio Ir/Ia (Ref 134). The current peak, Ir, for the reactivation scan is normalized for the grain- boundary area, while the initial anodic current peak remains normalized to As.

(Eq 34)

ir/ia ratios approaching 1 imply sensitization. A number of investigators have correlated this electrochemically derived ratio with optical metallographic evaluations of the degree of material sensitization such as those outlined in ASTM A 262 (Ref 135). This has been accomplished for several different Fe-Ni-Cr alloys (Ref 136, 137, 138). The technique is nondestructive to the underlying metal and can be applied in the field. Few methods exist for determination of intergranular corrosion rates. Serial metallographic sectioning and the foil-penetration method are the most prominent. In the foil-penetration rate, intergranular penetration is sensed using a foil held at open circuit or under potentiostatic control. The method can be repeated at several foil

thicknesses to determine whether rates are linear or not. In this way, the rate of penetration on the fastest intergranular path can be determined (Ref 139, 140).

References cited in this section 129.

P. Novak, R. Stefec, and F. Franz, Corrosion, Vol 31 (No. 10), 1975, p 344

130. W.L. Clarke, V.M. Romero, and J.C. Danko, Paper (preprint 180), Proc. Corrosion Conference, 1977, National Association of Corrosion Engineers, 1977 131. W.L. Clarke, R.L. Cowan, and W.L. Walker, Intergranular Corrosion of Stainless Alloys, STP 656, R.F. Steigerwald, Ed., American Society for Testing and Materials, 1978, p 99 132. “Test Method for Electrochemical Reactivation (EPR) Test Method for Detecting Sensitization of AISI Type 304 and 304L Stainless Steels,” G 108, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 133.

M. Akashi, T. Kawamoto, and F. Umemura, Corros. Eng., Vol 29, 1980, p 163

134.

A.P. Majidi and M.A. Streicher, Corrosion, Vol 40 (No. 11), 1984, p 584

135. “Practice for Detecting Susceptibility to Intergranular Attack in Wrought Nickel- Rich, Chromium Bearing Alloys,” A 262, Steel—Plate, Sheet, Strip, Wire; Stainless Steel Bar, Vol. 01.03, Annual Book of ASTM Standards, ASTM 136.

J.B. Lee, Corrosion, Vol 42 (No. 2), 1986, p 67

137.

A. Roelandt and J. Vereecken, Corrosion, Vol 42 (No. 5), 1986, p 289

138.

J.R. Scully and R. Kelly, Corrosion, Vol 42 (No. 9), 1986, p 537

139.

F. Hunkeler and H. Bohni, Corrosion, Vol 37, 1981, p 645

140.

A. Rota and H. Bohni, Werkst. Korros, Vol 40, 1989, p 219


Electrochemical Methods for Evaluation of Corrosion Rates under Paints Electrochemical impedance spectroscopy (EIS) techniques offer an advanced method of evaluating the performance of metallic-based coatings (passive film forming and/or conversion coatings) and the barrier properties of organic coatings and corrosion rates under paints (Ref 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151). The method does not accelerate the corrosion reaction and is nondestructive. The technique is quite sensitive to changes in the resistive-capacitive nature of coatings as well as the electrochemical interface. It is possible to monitor the polarization resistance of the corroding interface with this technique. In this respect,

the electrochemical impedance technique offers several advantages over direct current (dc) electrochemical techniques in that resistances related to the corrosion rate can be separated from the high dc resistance of the dielectric coating. This is not possible with the dc methods. Because a large frequency bandwidth is used for the applied signal (usually from the mHz range to the kHz range), the electrochemical impedance technique is “spectroscopic” and surpasses the capabilities of single-frequency impedance methods. The reason for this lies in the capability of the electrochemical impedance technique to discriminate between the resistive properties of the coating because of its ionic and/or electronic conductivity, the capacitive nature of the coating due to its dielectric constant, area, and thickness and the RC time constant of the electrified interface. Although impedance circuit models for coatings (Fig. 11) contain more elements than the model shown in Fig. 4, frequency regimes in which impedance information is primarily due to the coating capacitance or coating resistance can be separated from one another and analyzed independently by using a broad frequency bandwidth. The barrier properties of the coating can be followed as a function of exposure time. Large decreases signify permeation of ionic species through the coating or the presence of defects in the coating (Ref 148, 149). However, this is not the main focus of this article, which addresses methods to determine corrosion rates. The polarization resistance may be determined from the low-frequency time constant as shown in Fig. 11. However, recall that conversion of Rp to corrosion rate requires that the area specific polarization resistance be known as discussed previously. Thus, a major problem is knowledge of the corroding area.

Fig. 11 Electrical equivalent circuit to simulate a coated steel panel with a defect. RD is the coating defect resistance, RT is a charge-transfer resistance (similar to Rp) at the metal interface where water has penetrated. Cdl is the double-layer capacitance; Cc is the coating capacitance. Source: Ref 144, 145, 147 There are at least two situations that need to be discussed. These are cases where the coating is defect free but acts as an electrolyte allowing underpaint corrosion and the case where there are discrete defects and corrosion only occurs at these sites. Consider the case of an excellent barrier coating with isolated defects. One of the key difficulties in determining the rates of corrosion under paints is assessment of the defect area that is actively corroding. There are several ways to determine active area. Extraction of capacitance from EIS data is possible from equivalent circuit model fitting for sufficiently degraded defective coatings, as shown in Fig. 12. However, conversion of capacitance to area requires knowledge of the area-specific capacitance at the interface between electrolyte and metal. Such area-specific capacitance values have been developed for only a few corroding situations (Ref 48). For instance, area-specific capacitance has been developed for steel in soil, but has not been developed for coatings (Ref 48). The breakpoint frequency is a useful method for estimating the area fraction of physical defects in an organic coating (Ref 147, 148, 149, 150). The dependency between the high-frequency breakpoint frequency and defect area can be described by: (Eq 35)

where Ad is the defect area where electrochemical reactions operate, A is the total painted surface area, and ρ is the resistivity of the coating at the defect.

Fig. 12 Theoretical Bode magnitude and phase-angle plots for various known electrochemically active defect areas for a coating containing a cylindrical defect penetrating the metal substrate and no delaminated regions. (a) Bode magnitude. (b) Bode phase angle. ASTM visual ratings according to standards D 610 and D 714 are included for comparison. Source: Ref 150

Equation 35 provides a rough measure of the estimated change in defect area. Increases in fbpt occur mainly as a function of the increases in the defective area. However, circumstances may exist where small increases in ε and very large decreases in ρ occur simultaneously so that fbpt does not relate linearly to defect area. Moreover, bare metal under delaminated coating regions may not be detected under all circumstances especially if the coating resistivity over the delaminated area and dielectric constant are identical to those in regions where the coating is not delaminated. The hypothetical relationship between fbpt and open defect area over a range of pore dimensions is shown in Fig. 12 (Ref 150). From knowledge of defect area, the corrosion rate could be determined at the pore sites. Electrochemical Noise Methods for Organic Coating Evaluation. Electrochemical noise methods have also been explored as a method to analyze the degradation of polymer- coated metals (Ref 151, 152, 153, 154). A variety of methods have been used. A common method involves the use of a cell with two identical working electrodes connected through a ZRA. The entire galvanic couple is coupled through a high-impedance voltmeter to an ideally noiseless reference electrode. Two samples often experience drastically different behavior; for instance, if one electrode has a coating defect and the other does not then asymmetrical electrode behavior results. The spectral noise impedance obtained is equivalent to the geometric mean of the moduli of individual electrode impedances. Instrument noise should be carefully considered when this approach is used to analyze highresistance coatings. In principle, a noise-resistance value equivalent to the polarization resistance can be obtained. Knowledge of the corroding area enables calculation of the corrosion rate in the areas where corrosion occurs. Methods have not been developed and consensus has not been reached on approaches for determining corrosion rates under paints using EN methods.

References cited in this section 48. J.N. Murray, P.J. Moran, and E. Gileadi, Corrosion, Vol 44 (No. 8), 1988, p 533 141. J.D. Scantlebury, K.N. Ho, and D.A. Eden, Electrochemical Corrosion Testing, STP 727, F. Mansfeld and U. Bertocci, Ed., American Society for Testing and Materials, 1981, p 187 142. 48

S. Narian, N. Bonanos, and M.G. Hocking, J. Oil Colour Chem. Assoc., Vol 66 (No. 2), 1983, p

143.

T.A. Strivens and C.C. Taylor, Mater. Chem., Vol 7, 1982, p 199

144.

F. Mansfeld, M.W. Kendig, and S. Tsao, Corrosion, Vol 38 (No. 9), 1982, p 478

145.

M. Kendig, F. Mansfeld, and S. Tsai, Corros. Sci., Vol 23 (No. 4), 1983, p 317

146.

R. Touhsaent and H. Leidheiser, Corrosion, Vol 28 (No. 12), 1982, p 435

147. S. Haruyama, M. Asari, and T. Tsuru, Corrosion Protection by Organic Coatings, The Electrochemical Society Proceedings Series, M. Kendig and H. Leidheiser, Jr., Ed., 1980, p 197 148.

J.R. Scully, J. Electrochem. Soc., Vol 136 (No. 4), 1989, p 979

149.

M.W. Kendig and J.R. Scully, Corrosion, Vol 46 (No. 1), 1990, p 22

150.

H.P. Hack and J.R. Scully, J. Electrochem. Soc., Vol 138 (No. 1), 1991, p 33–40

151.

F. Mansfeld and C.C. Lee, J. Electrochem. Soc., Vol 144, 1997, p 2068–2071

152. F. Mansfeld, L.T. Han, C.C. Lee, C. Chen., G. Zhang, and H. Xiao, Corros. Sci., Vol 39, 1997, p 255–279

153. D.E. Tallman and G.P. Bierwagen, Paper No. 380, Proc. Corrosion Conference 1998, National Association of Corrosion Engineers, 1998 154.

A. Aballe, A. Bautista, U. Bertocci, and F. Huet, Corrosion, Vol 57, 2001, p 35


Nonelectrochemical Methods That Determine Cumulative Mass Loss Corrosion Rate Determination through Gravimetric Determination of Mass Loss. Corrosion rates are often determined from gravimetric measurement of mass loss in the laboratory either under conditions that mimic those in the field or under conditions that are accelerated. The latter is often achieved by raising the temperature or concentrating corrodents in the aqueous solutions used for testing. Field exposure is also simulated by various lab cabinet methods such as salt spray, prohesion, and Q-fog methods that introduce various weathering cycles including full immersion. The typical wisdom is that the accelerated test protocol produces the same form of corrosion, accelerates the mechanism of corrosion that is operative in the field, and correlates in some manner with field lifetimes or rates. Often the latter is not achieved in accelerated lab tests. In these cases, a relative ranking of materials or treatments is gained from corrosion rate determination by mass loss. It is desired that the same ranking is obtained in field testing. Mass-loss methods can be attractive for use in weatheringtype environments that involve wetting and drying. Under these conditions, electrochemical methods may not be suitable. ASTM standards cover many important procedural factors such as surface finish, and postexposure cleaning required for successful determination of mass loss. For instance, postcleaning exposure can introduce errors if cleaning methods remove additional metal besides the corrosion products. Conversely, nonremoval corrosion products or other deposits can produce mass gain that either partially or completely offsets small mass losses. Ideally, posttest cleaning methods rapidly remove corrosion products, but corrode underlying metal only a negligible amount. It follows that it can be difficult to determine the corrosion rates of passive metals by massloss methods because only very small mass changes occur over substantial time periods. In order to obtain measurable mass loss after exposure of a specimen, a rule used is that the exposure period must equal or exceed the ratio 2000/mpy = the number of exposure hours necessary (Ref 4). Instantaneous corrosion rate determination is not possible from mass loss because the time period of corrosion is uncertain and may not equal the exposure time period. Consequently, corrosion rate determination from a single mass-loss measurement represents an average over the period of exposure. The average corrosion rate can be obtained simply from standard relationships based on Faraday's law given a known mass loss measured after specimen cleaning (Ref 4, 155): (Eq 36) where K is a constant (see Table 3), T is the time of exposure (hours), A is the area (cm2), W is mass loss (grams), and ρ is material density in g/cm3.

Table 3 Constants K used in Eq 36 for desired penetration law units and mass-loss rate units Corrosion rate units desired Constant K in corrosion rate equation 3.45 × 106 Mils per year (mpy) 3.45 × 103 Inches per year (ipy) 2.87 × 102 Inches per month (ipm) 8.76 × 104 Millimeters per year (mm/yr) 8.76 × 107 Micrometers per year (μm/yr) 2.78 × 106 Picometers per second (pm/s) 1.00 × 104 × ρ Grams per square meter per hour (g/m2 · h) 2.40 × 106 × ρ Milligrams per square decimeter per day (mdd) 2 Micrograms per square meter per second (μg/m · s) 2.78 × 106 × ρ Note: ρ is the density in g/cm3. Source: Ref 155 Many different units are used to express corrosion rates. Using various values of K, mass- loss rate in different units can be produced (Ref 155). Appropriate modification of Eq 36 can be used to produce mass-loss rates per unit area (Table 3). Corrosion usually does not occur at a constant rate during the entire period of exposure. Procedures exist to ascertain whether or not the corrosion rate is increasing, decreasing, or constant with time. Removal at unit time periods in the beginning or end of a total exposure period or at prescribed time intervals can be used to ascertain this information (Ref 4, 5). Periodic determination of mass loss can be converted to mass-loss rate and penetration rate, assuming uniform attack. Mass-loss information can be obtained from full immersion in aqueous environments, alternate immersion, splash and spray, or gaseous corrosion exposure. Moreover, massloss methods may be useful in extremely low conductivity environments where electrochemical methods find difficulty. Corrosion rate determination by mass-loss methods can be misleading when corrosion occurs in a highly localized manner. In these cases, a depth gage, calipers, microscopic methods, and metallographic cross sections are preferred. Corrosion Rate Determination via Electrical-Resistance Methods. The electrical-resistance test method operates on the principle that the electrical resistance of the corrosion measuring element depends on the electrical resistance of the metallic cross section. Resistance is given by: (Eq 37) where R is resistance of sensing element, ρ is resistivity of sensing element (temperature-dependent material property), l is length of sensing element, and A is cross-sectional area. The resistance of the sensing element increases as the cross-sectional area of the sensing element decreases. Thus, the resistance increase is proportional to the reduction in cross-sectional area of the resistance-sensing element and corrosion can be detected (Ref 156, 157). Clearly, this method is only suitable in cases of uniform corrosion since high aspect ratio pits may not decrease the resistance of the sensing element in proportion to locally high corrosion rates. In practice, a resistance-measuring element constructed of the same material as the material of interest is exposed in the test environment. One shortcoming is when the component of interest cannot be matched exactly in the choice of sensing element. Often a second sensing element protected from the corrosive environment is measured simultaneously. The ratio of the resistances changes is determined. This compensates for any effects of exposure temperature on resistivity. The measurement taken gives the cumulative metal loss at the time of the measurement. The first derivative of measurements taken over time can be used to determine the rate of change of cross-sectional area A and, consequently, mass-loss rate. In other words, the slope of the mass-loss curve versus time can yield the mass loss per unit area per unit time. The slope at any time gives the corrosion rate at that time. The method can be used in atmospheric, gaseous, nonaqueous, or other environments where electrochemical methods may be compromised. Corrosion Rate Determination Using Magnetic Methods. A magnetic field is created by current flow through a wire coil. The magnetic induction whose magnitude is the magnetic flux density can be enhanced by placing a ferromagnetic material in the center of such a wire coil. The saturation flux or magnetic induction reaches a saturation value upon application of a sufficient magnetic field. This saturation flux is proportional to the mass of the ferromagnetic material and its relative permeability. Mass loss by uniform corrosion can be detected by a

decrease in the saturation value upon application of a magnetic field strength that saturates after corrosion has occurred and mass has been lost. Thus, the change in magnetic saturation induction of a ferromagnetic material can be used to determine the loss in mass of the test object. As in the case of the electrical-resistance methods, localized corrosion will not provide a change in magnetic field strength proportional to very localized depth of attack. A series of measurements on a single electrode provides a means to determine the rate of mass loss and corrosion rate. This method has been applied to coated metals and is capable of determining corrosion rates under painted cobalt (Ref 158, 159). An additional means of using magnetic effects to measure corrosion rates has been pursued using superconducting quantum interference devices (SQUIDs) to measure the electrical currents within metals resulting from corrosion activity on the metal surface. Magnetic fields are generated perpendicular to the metal surface with an intensity that scales with the corrosion rate. The SQUID acts as a sort of noninvasive ammeter, measuring the electron movement associated with corrosion activity. Early work by Bellingham, et al. (Ref 160, 161) demonstrated that corrosion activity did cause detectable magnetic signals. More recently, Abedi (Ref 162) has shown that a correlation exists between the temporal summation of the spatially integrated magnetic activity (referred to as TSSIMA) and the mass loss due to uniform corrosion for aluminum alloy 7075-T6 (UNS A97075-T6) in 0.1 M NaOH. More recent work has shown a correlation between corrosion activity and TSSIMA for crevice corrosion as shown in Fig. 13. The spatial distribution of the damage on surface (Fig. 13) corresponds to the distribution of magnetic activity observed over the 190 h of exposure.

Fig. 13 Noninvasive detection of hidden corrosion. (a) Micrograph of the lower (hidden) surface on a planar 2024-T3 sample that was exposed to a corrosive solution typical of an aircraft lap joint. The dark gray area in the upper center of the image corresponds to noncorroded aluminum. (b) The cumulative magnetic activity of the sample showing the more active regions in white and less active ones in dark gray. Source: Ref 163 The potential advantages of SQUID measurements include their high sensitivity, their noninvasive nature, and thus their ability to probe corrosion activity within occluded regions for which direct electrochemical measurements are not possible (Ref 164). Quantitative interpretation of the signals in terms of metal dissolution still requires development of a stronger theoretical underpinning. In the absence of such, the measurement of corrosion rate requires extensive calibration studies and a control of specimen geometry and corrosion mode. Corrosion Rate Determination Using the Quartz Crystal Microbalance. The quartz crystal microbalance offers a method for determining the mass loss or gain on an electrode material (m) (Ref 45, 165). In this case, the electrode material must first be deposited on a substrate with piezoelectric properties such as a slice of quartz crystal. Hence, the corrosion rate of the material deposited can be determined. The quartz crystal will oscillate at a frequency, f0, given by its geometry when a sinusoidal electrical signal is applied at this frequency. The

frequency of oscillation is sensitive to the mass change on the crystal surface as expressed by the Sauerbrey equation (Ref 165): (Eq 38) where Δf is the frequency change given by the addition or subtraction of mass per unit area of quartz crystal, n is the harmonic number of oscillation, μ is the shear modulus of quartz, and ρ is the density of quartz. When these constants are considered together, a single constant known as the sensitivity factor, Cf, is produced such that the mass change is directly proportional to the frequency change. This method can be used to determine mass loss or gain over time in gas- phase, thin-layer, or solution-phase corrosion. Unfortunately, the behavior of the crystal depends on the medium in which a crystal is operating. Therefore, a coating, film, and/or aqueous solution change the values of f0 and Cf because such a medium effectively couples to the crystal surface and creates additional resistance to the shear-wave oscillation used to detect a mass change. The method is quite sensitive to mass change. For instance, almost 0.01 μg/cm2 mass change can be detected by a 1 Hz change in frequency. The measurement taken gives the cumulative metal loss at the time of the measurement. The first derivative of many measurements taken over time can be used to determine the rate of change of mass and, consequently, the corrosion rate over a known area. This method has been used in several corrosion studies (Ref 166, 167). The method has the advantage of being useful for corrosion-rate determination in the gas phase, in nonaqueous environments, or when other electrochemical reactions may occur at equal or greater rates as the metal dissolution rate. However, the method has the disadvantage of being unable to detect and properly ascertain the rate of localized corrosion when only highly localized changes in mass occur. The resulting frequency change will not reflect the locally high amounts of mass loss. Solution Analysis Methods. Another nonelectrochemical means of assessing corrosion rate is the analysis of solution composition (Ref 168). As a metallic material corrodes, the metal cations are released into solution. If they can be quantitatively captured and accurately analyzed, the total dissolution of the material can be calculated directly. The challenge in such measurements is related to distribution of the metal-dissolution products. To the extent that cations precipitate as corrosion products, the measured solution concentration will be lowered. However, recent work using solution analysis methods to monitor corrosivity of secondary waters in a nuclear power plant toward the steel piping (Ref 169) has demonstrated that measurement of highly soluble alloying elements may provide an alternative approach. The concentration of soluble manganese was monitored online by ion chromatography and shown to correlate to the amount of solid iron corrosion products collected. Although the manganese was present at substantially less than 1 wt% in the steel, its high solubility in the nearneutral pH of the secondary water allowed it to be monitored whereas the iron precipitated. Metrological methods, to some degree, represent the most direct approach to measuring corrosion rate. Establishing the extent of material thinning can be done using several approaches including x-ray measurements of thickness, optical measurements of surface topography, and scanning probe microscopies. Xray radiography measures the material density as a function of position. Calibration standards allow the radiograph to be converted to a thickness map. In the case of optical metrology measurements, a number of commercial methods allow high-resolution measurements of surface topography. Confocal laser scanning microscopy (CLSM), white light interferometry, and scanning laser profilometry systems all have advantages and disadvantages with respect to speed, lateral and depth resolution, field of view, dynamic range, and cost (Ref 170, 171, 172). The scanning probe microscopies (STM, AFM, and NSOM) have been used to measure extremely low corrosion rates (Ref 173). Assuming the availability of a datum, that is, a position of known thickness (or a calibration curve in the case of radiography), these methods allow quantitative measurements of the material lost due to corrosion. It must be noted that with the exception of painstaking serial, cross-sectional metallography, all metrology methods are line-of-sight for the radiation used. This situation leads to the need to assume that all damage intersects the surface. Undercutting attack, as seen in some pitting systems, can go undetected so the corrosion rate is underestimated. However, corrosion under organic paint can be observed since thin paint layers are transparent to laser radiation (Ref 174). Thin-layer activation involves the production of a radioactive layer of material on the surface of the structure of interest. Typically, this layer is made via the exposure of a small section to a beam of high-energy charged

particles. Gamma radiation is emitted from the sample as the isotope decays and is detected. As the surface corrodes, less radiation is emitted as the total mass is decreased. In this way, the corrosion rate of the structure can be monitored. The levels of radiation emitted are very small, but the existence of high-sensitivity radiation detectors makes this method highly sensitive. Low corrosion rates can be monitored in situ, and the effects of corrective actions determined quickly (Ref 175).

References cited in this section 4. D.A. Jones, Principles and Prevention of Corrosion, 2nd ed., Macmillan Publishing, 1996 5. M.G. Fontana and N.D. Greene, Corrosion Engineering, McGraw-Hill, 1978 45. A.J. Bard and L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications, 2nd ed., John Wiley & Sons, 2001 155. “Standard Practice for Preparing, Cleaning and Evaluating Corrosion Tests Specimens,” G 1, Wear and Erosion; Metal Corrosion, Vol. 03.02, Annual Book of Standards, ASTM 156. “Standard Guide for On-line Monitoring of Corrosion in Plant Equipment (Electrical and Electrochemical Methods),” G 96, Wear and Erosion; Metal Corrosion, Vol. 03.02, Annual Book of Standards, ASTM 157. G. Cooper, Sensing Probes and Instruments for Electrochemical and Electrical Resistance Corrosion Monitoring, STP 908, ASTM 158. D.B. Mitton, Ph.D. thesis, University of Manchester Institute of Science and Technology, Manchester, U.K., 1989 159. N.J. Cantini, D.B. Mitton, N. Eliaz, G. Leisk, S.L. Wallace, F. Bellcuci, G.E. Thompson, and R.M. Latanision, Electrochem. Solid State Lett., Vol 396, 2000, p 275–278 160. J.G. Bellingham, M.L.A. MacVicar, M. Nisenoff, and P.C. Searson, J. Electrochem. Soc., Vol 133, 1986, p 1753 161.

J.G. Bellingham, M.L.A. MacVicar, and M. Nisenoff, IEEE Trans. Magn., Vol 23, 1987, p 477

162.

A. Abedi, Ph.D. Disseration, Vanderbilt University, 2000

163. A. Abedi, J.J. Fellenstein, A.J. Lucas, and J.P. Wikswo, Jr., Rev. Sci. Instrum., Vol 70 (No. 12) , 1999, p 4640–4651 164. G. Skennerton, A. Abedi, R.G. Kelly, and J.P. Wikswo, Jr., J. Corros. Sci. Eng., Vol 3, Paper 2, 2000 165.

G. Sauerbrey, Z. Phys., Vol 155, 1959, p 206

166.

R. Schumacher, A. Muller, and W. Stockel, J. Electroanal. Chem., Vol 219, 19876, p 311

167.

D. Jope, J. Sell, H.W. Pickering, and K.G. Weil, J. Electrochem. Soc., Vol 142, 1995, p 2170

168.

C.S. Brossia and R.G. Kelly, Corrosion, Vol 54, 1998, p 145–154

169.

D. Bostic, G. Burns, and S. Harvey, J. Chromatogr., Vol 602, 1992, p 163

170.

M.A. Alodan and W.H. Smyrl, J. Electrochem. Soc., Vol 145, 1998, p 1571

171.

J.B. Pawley, Ed., Handbook of Biological Confocal Microscopy, 2nd ed., Plenum Press, 1995

172.

N. Llorca-Isern, M. Puig, and M. Español, J. Therm. Spray Technol., Vol 8, 1999, p 73

173.

L.F. Garfias-Mesias and W.H. Smyrl, J. Electrochem. Soc., Vol 146, 1999, p 2495

174. O. Schneider, G.O. Ilevbare, J.R. Scully, and R.G. Kelly, Electrochem. Solid State Sci. Lett., Vol 4, 2001, p B35 175.

D.E. Williams and J. Asher, Corros. Sci., Vol 24, 1984, p 185


References 1. J.O'M. Bockris and A.K.N. Reddy, Modern Electrochemistry-2, Plenum Press, 1970 2. E. Gileadi, Electrode Kinetics for Chemists, Chemical Engineers and Materials Scientists, VCH Publishers, 1993 3. N.D. Tomashov, Theory of Corrosion and Protection of Metals, Macmillan Publishing, 1966 4. D.A. Jones, Principles and Prevention of Corrosion, 2nd ed., Macmillan Publishing, 1996 5. M.G. Fontana and N.D. Greene, Corrosion Engineering, McGraw-Hill, 1978 6. J. Newman, Electrochemical Systems, Prentice-Hall, 1973 7. H.H. Uhlig and R.W. Revie, Corrosion and Corrosion Control, John Wiley & Sons, 1985 8. J.O'M. Bockris, B.E. Conway, E. Yeager, and R.E. White, Ed., Electrochemical Materials Science, Vol 4, Comprehensive Treatise of Electrochemistry, Plenum Press, 1981 9. R. Baboian, Ed., Electrochemical Techniques for Corrosion, National Association of Corrosion Engineers, 1977 10. U. Bertocci and F. Mansfeld, Ed., Electrochemical Corrosion Testing, STP 727, American Society for Testing and Materials, 1979 11. Laboratory Corrosion Tests and Standards, STP 866, G. Haynes and R. Baboian, Ed., American Society for Testing and Materials, 1985

12. Corrosion Monitoring in Industrial Plants Using Nondestructive Testing and Electrochemical Methods, STP 908, G.C. Moran and P. Labine, Ed., American Society for Testing and Materials, 1986 13. R. Baboian, Ed., Electrochemical Techniques for Corrosion Engineers, National Association of Corrosion Engineers, 1986 14. R. Baboian, W.D. France, Jr., L.C. Rowe, and J.F. Rynewicz, Ed., Galvanic and Pitting CorrosionField and Laboratory Studies, STP 576, American Society for Testing and Materials, 1974 15. B. Poulson, Corros. Sci., Vol 23 (No. 4), 1983, p 391–430 16. D.D. MacDonald, Transient Techniques in Electrochemistry, Plenum Press, 1977 17. C. Wagner and W. Traud, Z. Electrochem., Vol 44, 1938, p 391 18. W.D. France, Jr., Controlled Potential Corrosion Tests: Their Application and Limitations, Mater. Res. Standard., Vol 9 (No. 8), 1969, p 21 19. “Standard Practice for Conventions Applicable to Electrochemical Measurements in Corrosion Testing,” G 3, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 20. “Standard Reference Test Method for Making Potentiostatic and Potentiodynamic Anodic Polarization Measurements,” G 5, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 21. “Practice for Conducting Potentiodynamic Polarization Resistance Measurements,” G 59, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 22. “Practice for Calculation of Corrosion Rates and Related Information from Electrochemical Measurements,” G 102, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 23. Z. Tafel, Phys. Chem., Vol 50, 1904, p 641 24. D. Britz, J. Electroanal. Chem., Vol 88, 1978, p 309 25. M. Hayes and J. Kuhn, J. Power Sources, Vol 2, 1977–1978, p 121 26. F. Mansfeld, Corrosion, Vol 38 (No. 10), 1982, p 556 27. L.L. Scribner and S.R. Taylor, Ed., The Measurement and Correction of Electrolyte Resistance in Electrochemical Tests, STP 1056, American Society for Testing and Materials, 1990 28. M. Stern and A.L. Geary, J. Electrochem. Soc., Vol 104, 1957, p 56 29. S. Evans and E.L. Koehler, J. Electrochem. Soc., Vol 108, 1961, p 509 30. M. Stern and R.M. Roth, J. Electrochem. Soc., Vol l04, 1957, p 390 31. M. Stern and A.L. Geary, J. Electrochem. Soc., Vol 105, 1958, p 638 32. M. Stern and E.D. Weisert, Experimental Observations on the Relation between Polarization Resistance and Corrosion Rate, Proc. ASTM, Vol 59, 1959, p 1280

33. “Standard Guide for On-Line Monitoring of Corrosion in Plant Equipment (Electrical and Electrochemical Methods),” G 96, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 34. S. Barnartt, Electrochim. Acta, Vol 15, 1970, p 1313 35. K.B. Oldham and F. Mansfeld, Corros. Sci., Vol 13, 1973, p 813 36. F. Mansfeld, J. Electrochem. Soc., Vol 120, 1973, p 515 37. F. Mansfeld and M. Kendig, Corrosion, Vol 37 (No. 9), 1981, p 556 38. R. Bandy and D.A. Jones, Corrosion, Vol 32, 1976, p 126 39. M.J. Danielson, Corrosion, Vol 36 (No. 4), 1980, p 174 40. J.C. Reeve and G. Bech-Nielsen, Corros. Sci., Vol 13, 1973, p 351 41. I. Epelboin, C. Gabrielli, M. Keddam, and H. Takenouti, Electrochemical Corrosion Testing, STP 727, F. Mansfeld and U. Bertocci, Ed., American Society for Testing and Materials, 1981, p 150 42. A.C. Makrides, Corrosion, Vol 29 (No. 9), 1973, p 148 43. D.D. MacDonald and M.C.H. McKubre, Electrochemical Corrosion Testing, STP 727, F. Mansfeld and U. Bertocci, Ed., American Society for Testing and Materials, 1981, p 110 44. J.R. Scully, Corrosion, Vol 56 (No. 2), 2000, p 199–218 45. A.J. Bard and L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications, 2nd ed., John Wiley & Sons, 2001 46. F. Mansfeld, Corrosion, Vol 36 (No. 5), 1981, p 301 47. “Practice for Verification of Algorithm and Equipment for Electrochemical Impedance Measurements,” G 106, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 48. J.N. Murray, P.J. Moran, and E. Gileadi, Corrosion, Vol 44 (No. 8), 1988, p 533 49. D.R. Franceschetti and J.R. Macdonald, J. Electroanal. Chem., Vol 101, 1979, p 307 50. B.A. Boukamp, Solid State Ionics, Vol 20, 1986, p 31–44 51. C.H. Hsu and F. Mansfeld, Corrosion, Vol 57, 2001, p 747 52. S.F. Mertens, C. Xhoffer, B.C. De Cooman, and E. Temmerman, Corrosion, Vol 53, 1997, p 381 53. J. Devay and L. Meszaros, Acta Chim. Acad. Sci. Hung., Vol 100, 1979, p 183 54. J.S. Gill, M. Callow, and J.D. Scantlebury, Corrosion, Vol 39, 1983, p 61 55. G.P. Rao and A.K. Mishra, J. Electroanal. Chem., Vol 77, 1977, p 121 56. L. Meszaros and J. Devay, Acta Chim. Acad. Sci. Hung., Vol 105, 1980, p 1

57. M.C.H. McKubre and B.C. Syrett, Corrosion Monitoring in Industrial Plants Using Nondestructive Methods, STP 908, G.C. Moran and P. Labine, Ed., American Society for Testing and Materials, 1986, p 433 58. R.W. Bosch, J. Hubrecht, W.F. Bogaerts, and B.C. Syrett, Corrosion, Vol 57, 2001, p 60–70 59. D.A. Eden, A.N. Rothwell, and J.L. Dawson, Paper 444, Proc. Corrosion Conference, 1991, National Association of Corrosion Engineers, 1991 60. J.L. Dawson, Electrochemical Noise Measurement for Corrosion Applications, STP 1277, J. Kearns, J.R. Scully, P.R. Roberge, D.L. Reichert, and J.L. Dawson, Ed., American Society of Testing and Materials, 1996, p 3–35 61. D.A. Eden, K. Hladky, D.G. John, and J.L. Dawson, Paper 276, Proc. Corrosion Conference, 1986, National Association of Corrosion Engineers, 1986 62. D.A. Eden and A.N. Rothwell, Paper 292, Proc. Corrosion Conference, 1992, National Association of Corrosion Engineers, 1992 63. A.N. Rothwell and D.A. Eden, Paper 223, Proc. Corrosion Conference, 1992, National Association of Corrosion Engineers, 1992 64. U. Bertocci, Electrochemical Noise Measurement for Corrosion Applications, STP 1277, J. Kearns, J.R. Scully, P.R. Roberge, D.L. Reichert, and J.L. Dawson, Ed., American Society of Testing and Materials, 1996, p 39–58 65. U. Bertocci and F. Huet, Corrosion, Vol 51, 1995, p 131 66. D.L. Reichert, Electrochemical Noise Measurement for Corrosion Applications, STP 1277, J. Kearns, J.R. Scully, P.R. Roberge, D.L. Reichert, and J.L. Dawson, Ed., American Society of Testing and Materials, 1996, p 79–89 67. P.C. Searson and J.L. Dawson, J. Electrochem. Soc., Vol 135 (No. 8), 1988, p 1908 68. F. Mansfeld and H. Xiao, Electrochemical Noise Measurement for Corrosion Applications, STP 1277, J. Kearns, J.R. Scully, P.R. Roberge, D.L. Reichert, and J.L. Dawson, Ed., American Society of Testing and Materials, 1996, p 59–78 69. G.P. Bierwagen, J. Electrochem. Soc., Vol 141, 1994, p L155 70. F. Mansfeld and H. Xiao, J. Electrochem. Soc., Vol 141, 1994, p 1403 71. F. Huet, J. Electrochem. Soc., Vol 142, 1995, p 2861 72. J.F. Chen and W.F. Bogaerts, Corros. Sci., Vol 37, 1995, p 1839 73. H. Xiao and F. Mansfeld, J. Electrochem. Soc., Vol 141, 1994, p 2332 74. F. Mansfeld and H. Xiao, J. Electrochem. Soc., Vol 140, 1993, p 2205 75. F. Mansfeld, L.T. Han, and C.C. Lee, J. Electrochem. Soc., Vol 143, 1996, p L286 76. U. Bertocci, C. Gabrielli, F. Huet, and M. Keddam, J. Electrochem. Soc., Vol 144, 1997, p 31

77. U. Bertocci, C. Gabrielli, F. Huet, and M. Keddam, J. Electrochem. Soc., Vol 144, 1997, p 37 78. A. Bautista and F. Huet, J. Electrochem. Soc., Vol 146, 1999, p 1730–1736 79. H. Hack and J.R. Scully, Corrosion, Vol 42 (No. 2), 1986, p 79 80. R. Baboian, Galvanic and Pitting Corrosion-Field and Laboratory Studies, STP 576, R. Baboian et al., Ed., American Society for Testing and Materials, 1974, p 5 81. F. Mansfeld and J.V. Kenkel, Galvanic and Pitting Corrosion-Field and Laboratory Studies, STP 576, R. Baboian et al., Ed., American Society for Testing and Materials, 1974, p 20 82. R.G. Kasper and C.R. Crowe, Comparison of Localized Ionic Currents as Measured from 1-D and 3-D Vibrating Probes, Galvanic Corrosion, STP 978, H.P. Hack, Ed., American Society for Testing and Materials, 1988, p 118 83. R.G. Kasper and C.R. Crowe, Comparison of Localized Ionic Currents as Measured from 1-D and 3-D Vibrating Probes, J. Electrochem. Soc., Vol 33, 1986, p 879 84. H.S. Isaacs, The Measurement of Galvanic Corrosion of Soldered Copper Using the Scanning Vibrating Electrode Technique, Corros. Sci., Vol 28, 1988, p 547 85. V.S. Voruganti, H.B. Huft, D. Degeer, and S.A. Bradford, Scanning Reference Electrode Technique for the Investigation of Preferential Corrosion of Weldments in Offshore Applications, Corrosion, Vol 47, 1991, p 343 86. Z. Szklarska-Smialowska, Pitting Corrosion of Metals, National Association of Corrosion Engineers, 1986 87. B.E. Wilde, Corrosion, Vol 28, 1972, p 283 88. B.C. Syrett, Corrosion, Vol 33, 1977, p 221 89. “Test Method for Conducting Cyclic Potentiodynamic Polarization Measurements for Localized Corrosion Susceptibility of Iron-, Nickel-, of Cobalt-Based Alloys,” G 61, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 90. N. Pessall and C. Liu, Electrochim. Acta, Vol 16, 1971, p 1987 91. B.E. Wilde and E. Williams, J. Electrochem. Soc., Vol 118, 1971, p 1057 92. N.G. Thompson and B.C. Syrett, Corrosion, Vol 48 (No. 8), 1992, p 649 93. “Standard Method for Conducting Cyclic Galvanostaircase Polarization,” G 100, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 94. “Test Method for Pitting or Crevice Corrosion of Metallic Surgical Implant Materials,” F 746, Annual Book of ASTM Standards, ASTM 95. B.E. Wilde and E. Williams, Electrochem. Acta, Vol 16, 1971, p 1671 96. K.K. Starr, E.D. Verink, Jr., and M. Pourbaix, Corrosion, Vol 2, 1976, p 47 97. N. Sridhar and G.A. Cragnolino, Corrosion, Vol 49, 1993, p 885

98. M. Yasuda, F. Weinburg, and D. Tromans, J. Electrochem. Soc., Vol 137, 1990, p 3708 99. S. Tsujikawa, T. Hasamatsu, and G. Boshoku, Corros. Eng., Vol 29, 1980, p 37 100.

T. Shinohara, S. Tsujikawa, and G. Boshoku, Corros. Eng., Vol 39, 1990, p 238

101.

S. Tsujikawa, Z. Heng, Y. Hisamatsu, and G. Boshoku, Corros. Eng., Vol 32, 1983, p 149

102. D. Dunn and N. Sridhar, Critical Factors in Localized Corrosion II, Proc. Electrochemical Society 95-15, P.M. Natishan, R.G. Kelly, G.S. Frankel, R.C. Newman, Ed., 1996, p 79–90 103. G.T. Gaudet, W.T. Mo, T.A. Hatton, J.W. Tester, J. Tilly, H.S. Isaacs, and R.C. Newman, Am. Inst., Chem. Eng. J., Vol 32 (No. 60), 1986, p 949 104.

B.A. Kehler, G.O. Ilevbare, and J.R. Scully, Corrosion, Vol 57 (No. 12), 2001, p 1042–1065

105.

P.C. Pistorius and G.T. Burstein, Philos. Trans. R. Soc. London, Vol 341, 1992, p 531

106.

T. Shibata, Trans. ISIJ, Vol 23, 1983, p 785

107.

T. Shibata and T. Takeyama, Corrosion, Vol 33, 1977, p 243

108.

T. Shibata and Y.-C. Zhu, Corros. Sci., Vol 36 (No. 1), 1994, p 153–163

109.

S.T. Pride, J.R. Scully, and J.L. Hudson, J. Electrochem. Soc., Vol 141, 1994, p 3028

110.

G.O. Ilevbare, J.R. Scully, J. Yuan, and R.G. Kelly, Corrosion, Vol 56 (No. 3), 2000, p 227–242

111.

J.E. Sweitzer, G.J. Shiflet, and J.R. Scully, Tri-Service Military Corrosion Conf. Proc., 2002

112. D.D. Macdonald, Critical Factors in Localized Corrosion, The Electrochemical Society Proceedings Series, PV 92-9, G.S.Frankel and R.C. Newman, Ed., 1992, p 144 113.

T. Suter and R.C. Alkire, J. Electrochem. Soc., Vol 148 (No. 1), 2001, p B36–B42

114.

G.T. Burstein and G.O. Ilevbare, Corros. Sci., Vol 38, 1996, p 2257–2265

115.


116.


117. H.S. Bertocci, Proc. Second International Conf. Localized Corrosion, NACE-9, H. Isaacs, U. Bertocci, J. Kruger, and S. Smialowska, Ed., National Association of Corrosion Engineers, 1990 118.

H.S. Bertocci and J. and Kruger, Surf. Sci., Vol 101, 1980, p 608–618

119. D.A. Eden and A.N. Rothwell, “Electrochemical Noise Data: Analysis and Interpretation,” Paper 92, Proc. Corrosion Conference, 1992, National Association of Corrosion Engineers, 1992 120.

P.C. Searson and J.L. Dawson, J. Electrochem. Soc., Vol 135, 1988, p 1908–1915

121. D.E. Williams, J. Stewart, and B.H. Balkwill, Proc. Symposium on Critical Factors in Localized Corrosion, Vol 92-9, G.S. Frankel and R.C. Newman, Ed., Electrochemical Society, 1992, p 36

122. S.T. Pride, J.R. Scully, and J.L. Hudson, Electrochemical Noise Measurement for Corrosion Applications, STP 1277, J.R. Kearns and J.R. Scully, et al., Ed., American Society for Testing and Materials, 1996, p 307 123.

J.R. Galvele, J. Electrochem. Soc., Vol 123, 1976, p 464

124. S. Turgoose and R.A. Cottis, “Corrosion Testing Made Easy, Electrochemical Impedance and Noise” (Houston, TX), NACE International, 2000 125.

R.A. Cottis, Corrosion, Vol 57, 2001, p 265

126. C. Gabrielli, F. Huet, M. Keddam, and R. Oltra, Localized Corrosion as a Stochastic Process, NACE—9 Advances in Localized Corrosion, H. Isaacs, U. Bertocci, J. Kruger, and S. Smialowska, Ed., National Association of Corrosion Engineers, 1990, p 93–108 127. J.R. Scully, S.T. Pride, H.S. Scully, and J.L. Hudson, Critical Factors in Local Corrosion II, Proc. Electrochemical Society, Vol 95-15, R.G. Kelly et al., Ed., 1996, p 15 128.

R.A. Cottis and S. Turgoose, Mater. Sci. Forum, Vol 2, 1995, p 192–194

129.

P. Novak, R. Stefec, and F. Franz, Corrosion, Vol 31 (No. 10), 1975, p 344

130. W.L. Clarke, V.M. Romero, and J.C. Danko, Paper (preprint 180), Proc. Corrosion Conference, 1977, National Association of Corrosion Engineers, 1977 131. W.L. Clarke, R.L. Cowan, and W.L. Walker, Intergranular Corrosion of Stainless Alloys, STP 656, R.F. Steigerwald, Ed., American Society for Testing and Materials, 1978, p 99 132. “Test Method for Electrochemical Reactivation (EPR) Test Method for Detecting Sensitization of AISI Type 304 and 304L Stainless Steels,” G 108, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM 133.

M. Akashi, T. Kawamoto, and F. Umemura, Corros. Eng., Vol 29, 1980, p 163

134.

A.P. Majidi and M.A. Streicher, Corrosion, Vol 40 (No. 11), 1984, p 584

135. “Practice for Detecting Susceptibility to Intergranular Attack in Wrought Nickel- Rich, Chromium Bearing Alloys,” A 262, Steel—Plate, Sheet, Strip, Wire; Stainless Steel Bar, Vol. 01.03, Annual Book of ASTM Standards, ASTM 136.

J.B. Lee, Corrosion, Vol 42 (No. 2), 1986, p 67

137.

A. Roelandt and J. Vereecken, Corrosion, Vol 42 (No. 5), 1986, p 289

138.

J.R. Scully and R. Kelly, Corrosion, Vol 42 (No. 9), 1986, p 537

139.

F. Hunkeler and H. Bohni, Corrosion, Vol 37, 1981, p 645

140.

A. Rota and H. Bohni, Werkst. Korros, Vol 40, 1989, p 219

141. J.D. Scantlebury, K.N. Ho, and D.A. Eden, Electrochemical Corrosion Testing, STP 727, F. Mansfeld and U. Bertocci, Ed., American Society for Testing and Materials, 1981, p 187

142. 48

S. Narian, N. Bonanos, and M.G. Hocking, J. Oil Colour Chem. Assoc., Vol 66 (No. 2), 1983, p

143.

T.A. Strivens and C.C. Taylor, Mater. Chem., Vol 7, 1982, p 199

144.

F. Mansfeld, M.W. Kendig, and S. Tsao, Corrosion, Vol 38 (No. 9), 1982, p 478

145.

M. Kendig, F. Mansfeld, and S. Tsai, Corros. Sci., Vol 23 (No. 4), 1983, p 317

146.

R. Touhsaent and H. Leidheiser, Corrosion, Vol 28 (No. 12), 1982, p 435

147. S. Haruyama, M. Asari, and T. Tsuru, Corrosion Protection by Organic Coatings, The Electrochemical Society Proceedings Series, M. Kendig and H. Leidheiser, Jr., Ed., 1980, p 197 148.

J.R. Scully, J. Electrochem. Soc., Vol 136 (No. 4), 1989, p 979

149.

M.W. Kendig and J.R. Scully, Corrosion, Vol 46 (No. 1), 1990, p 22

150.

H.P. Hack and J.R. Scully, J. Electrochem. Soc., Vol 138 (No. 1), 1991, p 33–40

151.

F. Mansfeld and C.C. Lee, J. Electrochem. Soc., Vol 144, 1997, p 2068–2071

152. F. Mansfeld, L.T. Han, C.C. Lee, C. Chen., G. Zhang, and H. Xiao, Corros. Sci., Vol 39, 1997, p 255–279 153. D.E. Tallman and G.P. Bierwagen, Paper No. 380, Proc. Corrosion Conference 1998, National Association of Corrosion Engineers, 1998 154.

A. Aballe, A. Bautista, U. Bertocci, and F. Huet, Corrosion, Vol 57, 2001, p 35

155. “Standard Practice for Preparing, Cleaning and Evaluating Corrosion Tests Specimens,” G 1, Wear and Erosion; Metal Corrosion, Vol. 03.02, Annual Book of Standards, ASTM 156. “Standard Guide for On-line Monitoring of Corrosion in Plant Equipment (Electrical and Electrochemical Methods),” G 96, Wear and Erosion; Metal Corrosion, Vol. 03.02, Annual Book of Standards, ASTM 157. G. Cooper, Sensing Probes and Instruments for Electrochemical and Electrical Resistance Corrosion Monitoring, STP 908, ASTM 158. D.B. Mitton, Ph.D. thesis, University of Manchester Institute of Science and Technology, Manchester, U.K., 1989 159. N.J. Cantini, D.B. Mitton, N. Eliaz, G. Leisk, S.L. Wallace, F. Bellcuci, G.E. Thompson, and R.M. Latanision, Electrochem. Solid State Lett., Vol 396, 2000, p 275–278 160. J.G. Bellingham, M.L.A. MacVicar, M. Nisenoff, and P.C. Searson, J. Electrochem. Soc., Vol 133, 1986, p 1753 161.

J.G. Bellingham, M.L.A. MacVicar, and M. Nisenoff, IEEE Trans. Magn., Vol 23, 1987, p 477

162.

A. Abedi, Ph.D. Disseration, Vanderbilt University, 2000

163. A. Abedi, J.J. Fellenstein, A.J. Lucas, and J.P. Wikswo, Jr., Rev. Sci. Instrum., Vol 70 (No. 12) , 1999, p 4640–4651 164. G. Skennerton, A. Abedi, R.G. Kelly, and J.P. Wikswo, Jr., J. Corros. Sci. Eng., Vol 3, Paper 2, 2000 165.

G. Sauerbrey, Z. Phys., Vol 155, 1959, p 206

166.

R. Schumacher, A. Muller, and W. Stockel, J. Electroanal. Chem., Vol 219, 19876, p 311

167.

D. Jope, J. Sell, H.W. Pickering, and K.G. Weil, J. Electrochem. Soc., Vol 142, 1995, p 2170

168.

C.S. Brossia and R.G. Kelly, Corrosion, Vol 54, 1998, p 145–154

169.

D. Bostic, G. Burns, and S. Harvey, J. Chromatogr., Vol 602, 1992, p 163

170.

M.A. Alodan and W.H. Smyrl, J. Electrochem. Soc., Vol 145, 1998, p 1571

171.

J.B. Pawley, Ed., Handbook of Biological Confocal Microscopy, 2nd ed., Plenum Press, 1995

172.

N. Llorca-Isern, M. Puig, and M. Español, J. Therm. Spray Technol., Vol 8, 1999, p 73

173.

L.F. Garfias-Mesias and W.H. Smyrl, J. Electrochem. Soc., Vol 146, 1999, p 2495

174. O. Schneider, G.O. Ilevbare, J.R. Scully, and R.G. Kelly, Electrochem. Solid State Sci. Lett., Vol 4, 2001, p B35 175.

D.E. Williams and J. Asher, Corros. Sci., Vol 24, 1984, p 185

M. Danielewski, Introduction to Fundamentals of Corrosion in Gases, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 87–89

Introduction to Fundamentals of Corrosion in Gases Revised by Marek Danielewski, AGH University of Science and Technology, Cracow, Poland

Introduction ENGINEERING MATERIALS are subject to deterioration when exposed to high-temperature environments. Whether they survive or not in technological applications depends on how fast they react. The rate of corrosion varies widely; some intermetallics (β-NiAl) react extremely slowly. Some metals (Fe) oxidize very rapidly, whereas other metals (Cr, Co) react relatively slowly. From the chemical point of view, the gas- metal reactions represent a broad class of heterogeneous reactions. The composition and structure of the scales produced on metals are a key factor in their behavior in technical applications. Historically, corrosion in gases has been primarily a problem in combustion systems. Thus, the gas-metal reactions are usually referred to as oxidation in its broad chemical sense, whether the reaction is with pure oxygen, water, sulfur dioxide (SO2), or whatever the gas might be. The corrosion product (oxide layer) is termed scale. Being the corrosion product, the protective properties of scale decrease the reaction rate. The concepts and methods developed to understand gas-metal reactions can be used to describe any arbitrary gas-

solid reaction at high temperature; for example, oxidation of the silicone carbide. The high temperature corrosion is a highly technical challenge; the reason for this is that the efficiency of thermal processes and engines increases with operating temperature. Such high- temperature service is especially damaging to most metals because of the exponential increase of reaction rate with temperature. In most cases, corrosion resistance at high temperatures does not accompany the good mechanical properties of structural materials; therefore, protective coatings must be applied. Electrochemical principles are insufficient to understand the mechanism of oxidation. For gaseous reactions, a basic understanding of the diffusion processes is much more profitable. The first results of a high-temperature corrosion study (not yet defined as corrosion and even diffusion) were published in 1684 by Boyle in Experiments and Considerations about the Porosity of Bodies in Two Essays. In studying reactive diffusion in the Cu-S system, Boyle reported the observation of interaction between copper and sulfur through examination of metallographic cross sections. Electrochemistry and aqueous corrosion principles were developed at the beginning of the 19th century. In 1855 Fick formulated the basic principles of diffusion in solids. The systematic study of high-temperature oxidation began in the 1920s. In 1933 Wagner published his pioneering paper on gas corrosion of metals. The first journal devoted to corrosion in gases, Oxidation of Metals, was published in the 1960s. The following articles introduce the subject of gas corrosion to professional engineers and students. A brief summary of thermodynamic concepts is followed by an explanation of the defect structure of solid oxides and the effect of these defects on the rate of mass transport. Commonly observed kinetics of oxidation are described and related to the observed corrosion mechanisms, as illustrated in Fig. 1 of the next article, “Thermodynamics of Gaseous Corrosion.” In high temperature gaseous corrosion, the oxidant first adsorbs on the metal surface in molecular (physical adsorption) and ionic form (chemical adsorption), and it may also dissolve in metal. Oxide nucleates at favorable sites and most commonly grows laterally, due to surface diffusion, to form a complete thin film (scale). As the scale thickens, it provides a protective barrier to shield the metal from the gas. For scale growth, electrons must move through the film to reach the oxidant atoms adsorbed on the surface, and oxidant ions and/or metal ions must move through the scale barrier. Diffusion of the oxidant into the metal may result in internal oxidation. Growth and thermal stresses in the oxide scale may create microcracks and/or delaminate scale from the underlying metal. Stresses affect the diffusion process and modify the oxidation mechanism and very often cause scale spallation. Improved oxidation resistance can be achieved by developing better alloys, by applying protective coatings, and by altering the composition of the gas phase.

M. Danielewski, Introduction to Fundamentals of Corrosion in Gases, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 87–89 Introduction to Fundamentals of Corrosion in Gases Revised by Marek Danielewski, AGH University of Science and Technology, Cracow, Poland

Fundamental Data Essential to an understanding of gaseous corrosion are the crystal structure and the density (molar volume) of the oxide and of the metal on which the oxide builds. Both may affect growth stresses in the oxide. For hightemperature service, the melting points of the oxide and metal, their structure, and thermal expansion coefficients, which affect oxide adherence during heating and cooling, are all important to know. These data for pure metals can be found in Table 1 here and in Table 1 of next article, “Thermodynamics of Gaseous Corrosion.” Likewise, data for oxides are shown in Table 2 and in Table 2 of the next article. Table 1 Properties of pure metals

Metal

Type

Density, g/cm3

Aluminum (Al) Antimony (Sb) Arsenic (As) Barium (Ba) Beryllium (Be) Bismuth (Bi) Cadmium (Cd) Calcium (Ca) Cerium (Ce) Cesium (Cs) Chromium (Cr) Cobalt (Co) Copper (Cu) Gallium (Ga) Germanium (Ge) Gold (Au) Hafnium (Hf) Indium (In) Iridium (Ir) Iron (Fe) Lanthanum (La) Lead (Pb) Lithium (Li) Magnesium (Mg) Manganese (Mn) Molybdenum (Mo) Neodymium (Nd) Nickel (Ni) Niobium (Nb) Osmium (Os) Palladium (Pd) Platinum (Pt)

… … … … IF-1 foil grade … … Rolled … … As-swaged … Annealed … … … Rod … Cold-drawn … … … … Sand cast … Annealed

2.6989 6.618 5.72 3.66 1.844 9.8 8.64 1.54 6.7 1.89 7.19 8.8 8.96 5.91 5.3234 19.32 13.31 7.31 22.65 7.87 6.166 11.34 53 1.74 7.44 10.22

Coefficient of linear thermal expansion (CTE) at 1000 °C 10-6 °C 10-6 °F … … … … … … … … 18.4 33.12 … … … … 22(a) 39.6(a) … … … … … … … … 24.8 44.64 … … … … 16.7 30.06 … … … … … … 24 43.2 … … … … … … … … 45.7 82.26 6.5 11.7

… Typical Annealed Sample Annealed … CP grade, annealed … Annealed Annealed Annealed … … … … … Cold worked … … … … …

7 8.88 8.6 22.5 12.02 21.45

… … 8.52 … 13.9 …

… … 15.33 … 25.02 …

86 21.03 12.4 12.3 4.81 2.329 10.491 971 2.6 16.65 11.3 9.33 5.765 6.45 7.29

… 6.65 … … … 4.44 19(a) … … 6.96 14.9 … … … …

… 11.97 … … … 7.99 34.2(a) … … 12.53 26.82 … … … …

Potassium (K) Rhenium (Re) Rhodium (Rh) Ruthenium (Ru) Selenium (Se) Silicon (Si) Silver (Ag) Sodium (Na) Strontium (Sr) Tantalum (Ta) Thorium (Th) Thulium (Tm) Tin-alpha (Sn) Tin-gamma (Sn) Tin-beta (Sn)

… 4.5 Titanium (Ti) … 19.3 Tungsten (W) Cast 19.07 Uranium (U) Cold rolled 6.11 Vanadium (V) … 4.472 Yttrium (Y) … 7.1 Zinc (Zn) … 6.53 Zirconium (Zr) (a) CTE at 100 °C (212 °F). Source: Ref 2, 3, 4, 5, 6, 7

10.1 … … 10.9 … … …

18.18 … … 19.62 … … …

Table 2 Structures and properties of oxides Oxide

Mineral

Density, g/cm3

Aluminium oxide (Al2O3) Barium oxide (BaO) Barium peroxide (BaO2) Beryllium oxide (BeO) Bismuth oxide (Bi2O3) Bismuth tetraoxide (Bi2O4) Calcium oxide (CaO) Calcium peroxide (CaO2) Cerium (III) oxide (Ce2O3) Cerium (IV) oxide (CeO2) Cesium oxide (Cs2O) Cesium superoxide (CsO22) Chromium (II, III) oxide (Cr3O4) Chromium (III) oxide (Cr2O3) Chromium (IV) oxide (CrO2) Chromium (VI) oxide (CrO3) Cobalt (II) oxide (CoO) Cobalt (II, III) oxide (Co3O4) Cobalt (III) oxide (Co2O3) Copper (I) oxide (Cu2O) Copper (II) oxide (CuO) Gallium (III) oxide (Ga2O3) Gallium suboxide (Ga2O) Gold (III) oxide (Au2O3) Hafnium (II) oxide (HfO2) Hafnium oxide (HfO2) Indium (III) oxide (In2O3) Iridium (III) oxide (Ir2O3) Iridium (IV) oxide (IrO2) Iron (II) oxide (FeO) Iron (II, III) oxide (Fe3O4) Iron (III) oxide (Fe2O3)

… … … Bromellite … … … … … Cerianite … … …

3.96 5.72 4.96 3.01 8.9 5.6 3.34 2.9 6.2 7.65 4.65 3.77 6.1

Decomposition temperature °C °F … … … … Max 450 Max 842 … … … … … … … … 200 392 … … … … … … … … … …

Crystal structure

Eskolaite

5.22

…

…

…

4.89

400

752

Hexagonal— rhomohedral Tetragonal

…

2.7

250

482

Orthorhombic

… …

6.44 6.11

… 900

… 1652

Cubic Cubic

… Cuprite Tenorite … … … Hafnia … … … … … Magnetite Hematite

5.18 6 6.31 6 4.77 … … 9.68 7.18 … 11.7 6 5.17 5.25

895 1800 … … Min 800 … … … … 1000 1100 … … …

1643 3272 … … Min 1472 … … … … 1832 2012 … … …

… Cubic Monoclinic Rhombic … … Monoclinic Cubic Cubic … Tetragonal Cubic Cubic Hexagonal— rhombohedral

Rhombohedral Cubic or Hexagonal Tetragonal Hexagonal Monoclinic … Cubic … Cubic Cubic Hexagonal Tetragonal Cubic

Oxide

Mineral

Density, g/cm3

Iron (III) oxide (Fe2O3) Lanthanum oxide (La2O3) Lead (II) oxide (PbO) Lead (II,III,IV) oxide (Pb3O4) Lead (II,IV) oxide (Pb2O3) Lead (IV) oxide (PbO2) Magnesium oxide (MgO) Manganese (II) oxide (MnO) Manganese (II, III) oxide (Mn3O4) Manganese (III) oxide (Mn2O3) Manganese (IV) oxide (MnO2) Molybdenum (III) oxide (Mo2O3) Molybdenum (IV) oxide (MoO2) Molybdenum (VI) oxide (MoO3) Neodymium oxide (Nd2O3) Nickel (II) oxide (NiO) Nickel (III) oxide (Ni2O3) Niobium (II) oxide (NbO) Niobium (IV) oxide (NbO2) Niobium (V) oxide (Nb2O5) Rhenium (IV) oxide (ReO2) Rhenium (V) oxide (Re2O5) Rhenium (VI) oxide (ReO3) Rhenium (VII) oxide (Re2O7) Silver (I) oxide (Ag2O) Silver (II) oxide (AgO) Titanium (II) oxide (TiO) Titanium (III) oxide (Ti2O3) Titanium (III,V) oxide (Ti3O5) Titanium dioxide (TiO2) Titanium dioxide (TiO2) Titanium dioxide (TiO2) Tungsten (IV) oxide (WO2)

Maghemite … … … … … Periclase Manganosite

Crystal structure

4.88 6.51 9.64 8.92

Decomposition temperature °C °F … … … … … … … …

Cubic … Orthorombic Tetragonal

10.05 9.64 3.6 5.37

… … … …

… … … …

Monoclinic Tetragonal Cubic Cubic

Hausmannite 4.84

…

…

Tetragonal

…

5

…

…

Cubic

Pyrolusite

5.08

535

995

Tetragonal

…

…

…

…

…

…

6.47

1100

2012

Tetragonal

…

4.7

…

…

Rhombohedral

… Bunsenite … … … … … … … …

7.24 6.72 … 7.3 5.9 4.6 11.4 7 6.9 6.1

… … 600 … … … 900 … 400 …

… … 1112 … … … 1652 … 752 …

Hexagonal Cubic Cubic Cubic Tetragonal Orthorhombic Orthorhombic Tetragonal Cubic …

… … … …

7.2 7.5 4.95 4.486

200 Min 100 … …

392 Min 212 … …

…

4.24

…

…

Cubic Monoclinic Cubic Hexagonal— rhomohedral Monoclinic

Rutile Anatase Brookite …

4.25 3.89 4.14 10.8

… … … 1500– 1700 … … …

… … … 2732– 3092 … … …

Tungsten (VI) oxide (WO3) … … Vanadium (II) oxide (VO) Vanadium (III) oxide Karelianite (V2O3)

7.2 5.758 4.87

Tetragonal Tetragonal Orthorhombic Monoclinic … … …

Oxide

Mineral

Density, g/cm3

Vanadium (IV) oxide (VO2) Vanadium (V) oxide (V2O5) Yttrium oxide (Y2O3) Zinc oxide (ZnO) Zinc peroxide (ZnO2) Zirconium oxide (ZrO2) Source: Ref 2, 6

… … … … … Zirconia

4.339 3.35 5.03 5.66 1.57 5.68

Decomposition temperature °C °F … … 1800 3272 … … … … Min 150 Min 302 … …

Crystal structure

… Orthorhombic Cubic Hexagonal … Monoclinic

References cited in this section 2. CRC Handbook of Chemistry and Physics, R.C. Weast, Ed., 62nd ed., CRC Press, 1981 3. R.B. Ross, Metallic Materials Specification Handbook, 4th ed., R.B. Ross, Chapman & Hall, 1992 4. Properties and Selection: Nonferrous Alloys and Special-Purpose Materials, Vol 2, ASM Handbook, ASM International, 10th ed., 1990 5. The Metals Databook, Alok Nayer, McGraw- Hill, New York, 1997 6. CRC Handbook of Chemistry and Physics, D.R. Lide, Ed., 79th ed., CRC Press, 1998 7. ASM Ready Reference: Thermal Properties of Metals, ASM International, 2002

M. Danielewski, Introduction to Fundamentals of Corrosion in Gases, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 87–89 Introduction to Fundamentals of Corrosion in Gases Revised by Marek Danielewski, AGH University of Science and Technology, Cracow, Poland

References 1. P. Kofstad, Oxidation Mechanisms for Pure Metals in Single Oxidant Gases, High Temperature Corrosion, R.A. Rapp, Ed., NACE International, 1983, p 123–138 2. CRC Handbook of Chemistry and Physics, R.C. Weast, Ed., 62nd ed., CRC Press, 1981 3. R.B. Ross, Metallic Materials Specification Handbook, 4th ed., R.B. Ross, Chapman & Hall, 1992 4. Properties and Selection: Nonferrous Alloys and Special-Purpose Materials, Vol 2, ASM Handbook, ASM International, 10th ed., 1990 5. The Metals Databook, Alok Nayer, McGraw- Hill, New York, 1997 6. CRC Handbook of Chemistry and Physics, D.R. Lide, Ed., 79th ed., CRC Press, 1998

7. ASM Ready Reference: Thermal Properties of Metals, ASM International, 2002

Z. Grzesik, Thermodynamics of Gaseous Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 90–96

Thermodynamics of Gaseous Corrosion Zbigniew Grzesik, University of Mining and Metallurgy (Krakow, Poland)

Introduction METALS can react chemically when exposed to air or to other more aggressive gases. The reaction rate of some metals is so slow that they are virtually unattacked, but for others, the reaction can be violent. As with most chemical processes, elevated-temperature service is more severe because of the exponential increase in reaction rate with temperature. The most common reactant is oxygen in the air; therefore, all gas-metal reactions are usually referred to as oxidation, using the term in its broad chemical sense whether the reaction is with oxygen, water vapor, hydrogen sulfide (H2S), or whatever the gas might be. Throughout this article, the process is called oxidation, and the corrosion product is termed an oxide. Corrosion in gases differs from aqueous corrosion in that electrochemical principles do not help greatly in understanding the mechanism of oxidation. For gaseous reactions, a fundamental knowledge of the diffusion processes involved is much more useful. The principles of high-temperature oxidation began to be understood in the 1920s, whereas electrochemistry and aqueous corrosion principles were developed approximately 100 years earlier. The first journal devoted to corrosion in gases, Oxidation of Metals, began publication in 1970. This article addresses thermodynamic concepts; the commonly observed kinetics of oxidation are described in the article “Kinetics of Gaseous Corrosion Processes” in this Volume. The mechanisms of oxidation are shown schematically in Fig. 1. The gas is first adsorbed on the metal surface as atomic oxygen. Oxide nucleates at favorable sites and most commonly grows laterally to form a complete thin film. As the layer thickens, it provides a protective scale barrier to shield the metal from the gas. For scale growth, electrons must move through the oxide to reach the oxygen atoms adsorbed on the surface, and oxygen ions, metal ions, or both must move through the oxide barrier. Oxygen may also diffuse into the metal.

Fig. 1 Schematic of the principal phenomena taking place during the reaction of metals with oxygen. Source: Ref 1 Growth stresses in the scale may create cavities and microcracks in the scale, modifying the oxidation mechanism or even causing the oxide to fail to protect the metal from the gas. Improved oxidation resistance can be achieved by selection of suitable alloys for the given environment and by application of protective coatings.

Reference cited in this section 1. P. Kofstad, Oxidation Mechanisms for Pure Metals in Single Oxidant Gases, High Temperature Corrosion, R.A. Rapp, Ed., NACE International, 1983, p 123–138

Z. Grzesik, Thermodynamics of Gaseous Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 90–96 Thermodynamics of Gaseous Corrosion Zbigniew Grzesik, University of Mining and Metallurgy (Krakow, Poland)

Fundamental Data Essential to an understanding of the gaseous corrosion of a metal are the crystal structure and the molar volume of the metal on which the oxide builds, both of which may affect growth stresses in the oxide. For hightemperature service, the melting point of the metal and the structural changes that take place during heating and cooling, which affect oxide adherence, must be known. These data are presented in Table 1 for pure metals. For the oxides, structures, melting and boiling points, molar volume, and oxide/ metal volume ratio (PillingBedworth ratio) are shown in Table 2. Table 1 Structures and thermal properties of pure metals Metal

Structure(a)

Aluminum Antimony Arsenic

fcc rhom rhom

Barium Beryllium

bcc (α) hcp (β) bcc rhom hcp (α) fcc (β) bcc (γ) fcc (δ) bcc bcc bcc (α) hcp (β) fcc fcc (α) hcp (β) bcc hcp bcc (α) hcp (β) bcc ortho diamond fcc fcc (α) hcp (β) bcc

Bismuth Cadmium Calcium Cerium Cesium Chromium Cobalt Copper Dysprosium Erbium Europium Gadolinium Gallium Germanium Gold Hafnium

Transformation temperature °C °F … … … … … …

Volume change on cooling(b), % °C … … …

… 1250 … … … 448 … 726 … … … 417 … … 1381 … … … 1235 … … …

… 2282 … … … 838 … 1339 … … … 783 … … 2518 … … … 2255 … … …

… 1742 …

… 3168 …

Melting point cm3 1220.7 1167.3 1139

… … -2.2 … … … -0.4 … … … … … -0.3 … … -0.1 … … … -1.3 … …

°F 660.4 630.7 Sublimation 615 729 … 1290 271.4 321.1 … 839 … 798 28.64 1875 … 1495 1084.88 … 1412 1529 822 … 1312 29.78 937.4

Molar volume(c) in.3 10.00 0.610 18.18 1.109 12.97 0.791

1344 … 2354 520.5 610 … 1542 … 1468 83.55 3407 … 2723 1984.78 … 2573 2784 1512 … 2394 85.60 1719.3

39 4.88 4.99 21.31 13.01 25.9 … 20.70 … 70.25 7.23 6.67 6.70 7.12 19.00 18.98 18.45 28.98 19.90 20.16 11.80 13.63

2.380 0.298 0.304 1.300 0.793 1.581 … 1.263 … 4.287 0.441 0.407 0.408 0.434 1.159 1.158 1.126 1.768 1.214 1.230 0.720 0.832

… … …

1064.43 … 2231

1947.97 10.20 … 13.41 4048 …

0.622 0.818 …

Metal

Holmium Indium Iridium Iron

Lanthanum

Lead Lithium Lutetium Magnesium Manganese

Mercury Molybdenum Neodymium Nickel Niobium Osmium Palladium Platinum Plutonium

Structure(a)

hcp tetr fcc (α) bcc (γ) fcc (δ) bcc (α) hex (β) fcc (γ) bcc fcc (β) bcc hcp hcp (α) cubic (β) cubic (γ) tetr rhom bcc (α) hex (β) bcc fcc bcc hcp fcc fcc α, β, γ δ, δ′, ε

Potassium bcc Praseodymium (α) hex (β) bcc Rhenium hcp Rhodium fcc Rubidium bcc Ruthenium hcp Samarium (α) rhom (β) hcp (γ) bcc Scandium (α) hcp (β) bcc Selenium (γ) hex Silicon diamond fcc Silver fcc Sodium (β) bcc Strontium (α) fcc (β) bcc

Transformation temperature °C °F … … … … … … 912 1674 1394 2541 … … 330 626 865 … … … … … -193 -315 … … … … 710 1310 1079 1974 … … … … … … 863 1585 … … … … … … … … … … … … 120, 248, 210, 410, 315 599 452, 846, 480 896 … … 795 1463 … … … … … … … … … … 734 1353 922 1692 … … 1337 2439 … … 209 408 … …

Volume change on cooling(b), % °C … … … … 1.0 -0.52 … 0.5 -1.3 … … … … … -3.0 -0.0 … … … -0.1 … … … … … …

°F 1474 156.63 2447 … … 1538 … … 918 327.4 180.7 1663 650 … … 1244 -38.87 2610 … 1021 1453 2648 ~2700 1552 1769 …

cm3 2685 313.93 4437 … … 2800 … … 1684 621.3 357.3 3025 1202 … … 2271 -37.97 4730 … 1870 2647 4474 ~4890 2826 3216 …

…

640

1184

… … -0.5 … … … … … … … … … … …

63.2 … 931 3180 1963 38.89 2310 … … 1074 … 1541 217 1410

… -237 557 …

… … … …

961.9 97.82 … 768

… -395 1035 …

Melting point

Molar volume(c) in.3 18.75 1.144 15.76 0.962 8.57 0.523 7.10 0.433 7.26 0.443 7.54 0.460 22.60 1.379 22.44 1.369 23.27 1.420 18.35 1.119 12.99 0.793 17.78 1.085 13.99 0.854 7.35 0.449 7.63 0.466 7.62 0.465 14.81 0.904 9.39 0.573 20.58 1.256 21.21 1.294 6.59 0.402 10.84 0.661 8.42 0.514 8.85 0.540 9.10 0.555 α α 12.04 0.735

145.8 … 1708 5756 3565 102 4190 … … 1965 … 2806 423 2570

ε 14.48 45.72 20.80 21.22 8.85 8.29 55.79 8.17 20.00 20.46 20.32 15.04 … 16.42 12.05

ε 0.884 2.790 1.269 1.295 0.540 0.506 3.405 0.499 1.220 1.249 1.240 0.918 … 1.002 0.735

1763.4 208.08 … 1414

10.28 23.76 34 34.4

0.627 1.450 2.075 2.099

Metal

Structure(a)

Transformation Volume change Melting point Molar temperature on cooling(b), % volume(c) °C °F °C °F cm3 in.3 Tantalum bcc … … … 2996 5425 10.9 0.665 Tellurium hex … … … 449.5 841.1 20.46 1.249 Terbium (α) hcp 1289 2352 … … … 19.31 1.178 (β) bcc … … … 1356 2472.8 19.57 1.194 Thallium (α) hcp 230 446 … … … 17.21 1.050 (β) bcc … … … 303 577 … … Thorium (α) fcc 1345 2453 … … … 19.80 1.208 (β) bcc … … … 1755 3191 21.31 1.300 Thulium hcp … … … 1545 2813 18.12 1.106 Tin (β) bct 13.2 55.8 27 231.9 449.4 16.56 1.011 Titanium (α) hcp 882.5 1621 … … … 10.63 0.649 (β) bcc … … … 1668 3034 11.01 0.672 Tungsten bcc … … … 3410 6170 9.55 0.583 Uranium (α) ortho 661 1222 … … … 12.50 0.763 (β) complex 769 1416 -1.0 … … 13.00 0.793 tetr (γ) bcc … … -0.6 1900 3452 8.34 0.509 Vanadium bcc 1910 3470 … … … … … Ytterbium (β) fcc 7 45 0.1 819 1506 24.84 1.516 Yttrium (α) hcp 1478 2692 … … … 19.89 1.214 (β) bcc … … … 1522 2772 20.76 1.267 Zinc hcp … … … 420 788 9.17 0.559 Zirconium (α) hcp 862 1584 … … … 14.02 0.856 (β) bcc … … … 1852 3366 15.09 0.921 (a) fcc, face-centered cubic; rhom, rhombohedral; bcc, body-centered cubic; hcp, hexagonal close-packed; ortho, orthorhombic; tetr, tetragonal; hex, hexagonal; bct, body-centered tetragonal. (b) Volume change on cooling through crystallographic transformation. (c) Molar volume at 25 °C (77 °F) or at transition temperature for structures not stable at 25 °C (77 °F). Source: Ref 2 Table 2 Structures and thermal properties of selected oxides Oxide

α-Al2O3 γ-Al2O3 BaO BaO2 BeO CaO CaO2 CdO Ce2O3 CeO2 CoO Co2O3 Co3O4

Structure

D51 (corundum) (defect-spinel) B1 (NaCl) Tetragonal (CaC2) B4 (ZnS) B1 (NaCl) C11 (CaC2) B1 (NaCl) D52 (La2O3) C1 (CaF2) B1 (NaCl) Hexagonal H11 (spinel)

Melting point °C 2015 γ→α 1923 450

°F 3659 … 3493 842

Boiling or Molar volume Volume (a) ratio decomposition (d) 3 3 °C °F cm in. 2980 5396 25.7 1.568 1.28 … … 26.1 1.593 1.31 ~2000 ~3632 26.8 1.635 0.69 d.800 d.1472 34.1 2.081 0.87

2530 2580 … ~1400 1692 ~2600 1935 … →CoO

4586 4676 … ~2552 3078 ~4712 3515 … …

~3900 2850 d.275 d.900 … … … d.895 …

~7052 5162 d.527 d.1652 … … … d.1643 …

8.3 16.6 24.7 18.5 47.8 24.1 11.6 32.0 39.7

0.506 1.013 1.507 1.129 2.917 1.471 0.708 1.953 2.423

1.70 0.64 0.95 1.42 1.15 1.17 1.74 2.40 1.98

Oxide

°C 2435 …

°F 4415 …

Boiling decomposition (d) °C °F 4000 7232 d.400 d.752

400 1326 1235 2340 … 1420

752 2419 2255 4244 … 2588

650 … d.1800 … … …

1202 … d.3272 … … …

70.1 12.3 23.8 47.8 44.3 12.6

4.278 0.751 1.452 2.917 2.703 0.769

1565

2849

…

…

30.5

1.861

…

…

…

…

…

…

…

γ-Fe2O3

D57 cubic

1457

2655

…

…

31.5

1.922

Fe3O4

H11 (spinel)

…

…

d.1538

d.2800

44.7

2.728

…

…

…

…

…

…

…

Monoclinic Cubic Defect B10(SnO) D53(Sc2O3) C4(TiO2) C1(CaF2) D52 hexagonal C1 (CaF2) B1 (NaCl) B1(NaCl) C4 (TiO2) D53 (Sc2O3) H11 (spinel)

1900 2812 …

3452 5094 …

… ~5400 d.500

… ~9752 d.932

31.9 21.7 19.5

1.947 1.324 1.190

0.50 1.72 1.67 1.26 1.20 1.78 on iron 2.15 on iron 1.02 Fe3O4 2.22 on iron 2.10 on iron ~1.2 FeO 1.35 1.62 1.32

… … … 2315 ~1700 2800 … … … 1705

… … … 4199 ~3092 5072 … … … 1301

d.850 d.1100 d.350 4200 1200 3600 … d.535 d.1080 …

d.1562 d.2012 d.662 7592 2192 6512 … d.995 d.1976 …

38.7 19.1 40.6 50.0 14.8 11.3 13.0 17.3 35.1 47.1

2.362 1.166 2.478 3.051 0.903 0.690 0.793 1.056 2.142 2.874

1.23 2.23 0.45 1.10 0.57 0.80 1.77 2.37 2.40 2.14

795 Sublimation 1275 1460 ~1900 1990 … 888 … 870 … 489 … …

1463 2327

… …

… …

30.7 27.3

1.873 1.666

3.27 0.57

2660 ~3452 3614 … 1630 … 1598 … 912 … …

… … … d.350 … d.500 … d.550 … d.1000 d.1100

… … … d.662 … d.932 … d.1022 … d.1832 d.2012

59.5 46.5 11.2 28.8 23.4 75.3 14.1 14.2 62.0 19.1 31.0

3.631 2.838 0.683 1.757 1.428 4.595 0.860 0.867 3.783 1.166 1.892

2.74 1.13 1.70 3.42 1.28 1.37 1.59 1.56 0.56 2.16 1.87

Cr2O3 Cs2O Cs2O3 CuO Cu2O Dy2O3 Er2O3 FeO

Structure

D51 (αAl2O3) Hexagonal (CdCl2) Cubic (Th3P4) B26 monoclinic C3 cubic Cubic (Tl2O3) Cubic (Tl2O3) B1 (NaCl)

α-Fe2O3 D51 (hematite)

Ga2O3 HfO2 HgO In2O3 IrO2 K2O La2O3 Li2O MgO MnO MnO2 Mn2O3 αMn3O4 MoO3 Na2O Nb2O5 Nd2O3 NiO OsO2 PbO Pb3O4 PdO PtO Rb2O3 ReO2 Rh2O3

Orthorhombic C1 (CaF2) Monoclinic Hexagonal B1 (NaCl) C4 (TiO2) B10 tetragonal Tetragonal B17 tetragonal B17 (PdO) (Th3P4) Monoclinic D51 (α-Al2O3)

Melting point

or Molar volume Volume (a) ratio 3 3 cm in. 29.2 1.782 2.02 66.3 4.046 0.47

ααon ααon

Oxide

SiO SiO2 SnO SnO2 SrO Ta2O5 TeO2 ThO2 TiO TiO2 Ti2O3 Tl2O3 UO2 U3O8 VO2 V2O3 V2O5 WO2 β-WO3 W2O5

Structure

Cubic β cristobalite C9 B10 (PbO) C4 (TiO2) B1 (NaCl) Triclinic C4 (TiO2) C1 (CaF2) B1 (NaCl) C4 (rutile) D51 (α-Al2O3) D53 (Sc2O3) C1 (CaF2) Hexagonal C4 (TiO2) D51 (α-Al2O3) D87 orthorhombic C4 (TiO2) Orthorhombic Triclinic

Melting point °C ~1700 1713 … 1127 2430 1800 733 3050 1750 1830 … 717 2500 … 1967 1970 690

Boiling or Molar volume Volume (a) decomposition (d) ratio 3 3 °F °C °F cm in. ~3092 1880 3416 20.7 1.263 1.72 3115 2230 4046 25.9 1.581 2.15 … d.1080 d.1976 20.9 1.275 1.26 2061 … … 21.7 1.324 1.31 4406 ~3000 ~5432 22.0 1.343 0.65 3272 … … 53.9 3.289 2.47 1351 1245 2273 28.1 1.715 1.38 5522 4400 7952 26.8 1.635 1.35 3182 ~3000 ~5432 13.0 0.793 1.22 3326 ~2700 ~4892 18.8 1.147 1.76 … d.2130 d.3866 31.3 1.910 1.47 1323 d.875 d.1607 44.8 2.734 1.30 4532 … … 24.6 1.501 1.97 … d.1300 d.2372 101.5 6.194 2.71 3573 … … 19.1 1.166 2.29 3578 … … 30.8 1.879 1.85 1274 d.1750 d.3182 54.2 3.307 3.25

~1550 ~2822 ~1430 ~2606 17.8 1.086 1.87 1473 … … … 32.4 1.977 3.39 Sublimation, ~1562 ~1530 ~2786 29.8 1.819 3.12 ~850 Y2O3 D53 (Sc2O3) 2410 4370 … … 45.1 2.752 1.13 ZnO B4 (wurtzite) 1975 3587 … … 14.5 0.885 1.58 ZrO2 C43 monoclinic 2715 4919 … … 22.0 1.343 1.57 (a) Molar volume at 25 °C (77 °F) or at transition temperature for structures not stable at 25 °C (77 °F). Source: Ref 3 Thermodynamics plays a very important role in studying the gaseous corrosion of metallic materials. It is possible to determine several critical parameters of the oxidation reaction on the basis of thermodynamic data, namely, the temperature and oxidizing gas pressure in which particular chemical compounds can be formed, the phase sequence in the multilayered scale growing on metal, the equilibrium composition of the gas mixture, and partial pressures of volatile oxidation products. In spite of its name, thermodynamics does not inform on the reaction rate; it does not describe “the dynamics” of the systems. Thermodynamics determines the equilibrium state. The Gibbs energy change, which is the driving force of any chemical reaction, is not related to the reaction rate. Reaction rate is a kinetic problem and depends on the mechanism of the slowest step of the overall reaction process. On the other hand, the thermodynamic considerations of metal-oxidant equilibria and phase diagrams are invaluable tools for the interpretation of the oxidation mechanisms and processes. Some typical examples of such considerations are described in the following section of this article.

References cited in this section 2. Properties of Pure Metals, Properties and Selection: Nonferrous Alloys and Pure Metals, Vol 2, 9th ed., Metals Handbook, American Society for Metals, 1979, p 714–831 3. R.C. Weast, Ed., Physical Constants of Inorganic Compounds, Handbook of Chemistry and Physics, 65th ed., The Chemical Rubber Company, 1984, p B68–B161


Free Energy of Reaction The chemical reaction describing the oxidation process of a pure divalent metal, M, by an oxidizing gas, X2 (oxygen, sulfur, and others), may be written as: (Eq 1) where MaXb is the reaction product (oxide, sulfide, etc.). The fundamental criterion, which allows for the evaluation of whether, at a given temperature and oxidant gas pressure, the oxidation process of a determined metal may occur, is the sign of the Gibbs energy change, ΔG (Ref 4, 5). The Second Law of Thermodynamics describes ΔG as: ΔG = ΔH - TΔS

(Eq 2)

where ΔH denotes the enthalpy of reaction, T is the absolute temperature, and ΔS is the entropy change. For a spontaneous reaction at constant pressure and temperature, ΔG is negative (ΔG < 0). In the case of nonspontaneous reactions, ΔG is positive (ΔG > 0), and if ΔG = 0, the system is at equilibrium. The Gibbs energy change of the chemical reaction given by Eq 1 equals the sum of the chemical potentials, μ, of all components present in the system (M, X2, MaXb): (Eq 3) The chemical potential, μi, and the activity, ai, of a given i-component are interrelated, as follows: (Eq 4) where R is the gas constant, and is the chemical potential of a given i-component in the standard state. Generally, the activities of pure solid components (such as metals and oxides) are equal to unity, and therefore μM =

and

. The activity of gaseous component X2 can be approximated by its pressure: (Eq 5)

Thus, Eq 3 may be presented in the following form: (Eq 6) Because the sum of the standard chemical potentials is the standard Gibbs energy change, ΔG0, then the previous equation may be written as: (Eq 7) At equilibrium, where ΔG = 0, Eq 7 can be presented in the form: (Eq 8) In the previous equation, from the following relation:

denotes the dissociation pressure of the MaXb compound and may be calculated

(Eq 9) Thermodynamically, if the oxidant pressure is lower than the calculated value of the equilibrium dissociation pressure of the MaXb compound, then the metal, M, is not oxidized. In the opposite case, the spontaneous oxidation reaction may occur. It should be emphasized that the calculated value of dissociation pressure of a given compound is related to the equilibrium state. In practice, thermodynamically unstable oxides can be formed during high-temperature oxidation of metals by gases. An example of such unstable oxides is Wustite (FeO), which is unstable in air unless the temperature is higher than 570 °C (1060 °F). In fact, Wustite is a major component of the scale formed on steel beyond this temperature, and after rapid cooling to room temperature, it remains in the scale as a result of the extremely slow decomposition kinetics of FeO. See the FeO binary phase diagram (Fig. 2); the diagram shows that other scale components are Fe2O3 and Fe3O4.

Fig. 2 Iron-oxygen phase diagram

References cited in this section 4. O. Kubaschewski and B.E. Hopkins, Oxidation of Metals and Alloys, Butterworths, 1962 5. I. Barin, O. Knacke, and O. Kubaschewski, Thermochemical Properties of Inorganic Substances, Springer, 1977


Richardson-Jeffes Diagrams The determination of the standard Gibbs energy change of formation of oxides as well as the corresponding dissociation pressures of the oxides as a function of temperature—without any calculations—is very convenient when using Richardson-Jeffes diagrams (Gibbs energy-temperature diagrams) (Ref 6). Such a diagram, Fig. 3, is a plot of the standard Gibbs energies of formation of the oxides per mole of oxygen, O2 (e.g., 2Ni + O2 = 2NiO, or Al + O2 = Al2O3), versus temperature. Auxiliary outer scales in the upper, right, and bottom part of the plot are used for determination of conditions in which a given metal is oxidized by pure oxygen gas or oxidizing gas mixtures, CO/CO2 and H2/ H2O. The standard Gibbs energy change of formation of any oxide at a given temperature can be read directly on the ΔG0 scale, for example, the Gibbs energy of formation of SiO 2 at 1200 °C (2190 °F) equals approximately -600 kJ for one mole of SiO2

Fig. 3 Richardson-Jeffes diagram showing standard Gibbs free energy of formation as a function of temperature for metal oxide systems. Source: Ref 7 The dissociation pressure of a given oxide at constant temperature can be read on the partial pressure of oxygen, , scale from the intersection of this scale with a straight line, drawn from index point labeled “O” at the upper left corner of the diagram through the appropriate temperature point on the line, related to the formation of a corresponding oxide. By way of example, the dissociation pressure of SiO2 is approximately 1020 atm O2 at 1200 °C (2190 °F). Thus, silicon is not oxidized to SiO2 at partial pressures lower than 10-20 atm at this temperature. On the other hand, from a thermodynamical point of view, SiO2 can be formed from silicon at any pressure of oxygen greater than 10-20 atm at 1200 °C (2190 °F). The determination of the dissociation pressures of oxides is of considerable practical consequence in predicting which metallic materials are oxidized

at given conditions and, as a consequence, how a low partial pressure of oxygen should be used to prevent oxidation. However, very low partial pressures of oxygen in ambient gases ( < 10-6 atm) are, in practice, seldom realized by means of vacuum systems or by oxygen-purified noble gases (such as argon or helium). Generally, to obtain very low partial pressure of oxygen, oxidizing gas mixtures are used in which oxygen is one of the components, for example, CO2-CO-O2 or H2O-H2-O2. The essential chemical reactions for these gas mixtures can be written as follows: (Eq 10) (Eq 11) The corresponding equilibrium partial pressures of oxygen, expressed as:

, established at equilibrium state may be

(Eq 12)

(Eq 13) where

and

are the standard Gibbs energy changes of the reactions in Eq 10 and 11, respectively.

As can be seen from Eq 12 and 13, for constant ratios of or , the partial pressure of oxygen is also constant and does not depend significantly on the total pressure of the system. In other words, at a given temperature, can be easily controlled by controlling the ratio of or . Consequently, it is possible to choose such a ratio of corresponding gases that the partial pressure of oxygen reaches the value of the dissociation pressure of a given oxide. Such an equilibrium ratio can be read off the Richardson-Jeffes diagrams. For instance, to find a ratio for which the Si-SiO2-O2 system is at equilibrium at 1000 °C (1830 °F), a straight line should be drawn from the index point labeled “C” at the lefthand scale through the 1000 °C (1830 °F) point on the line related to the formation of SiO2. The equilibrium ratio can be read from the intersection of this line with the same way, appropriate equilibrium ratios of

scale and equals 106. In the

can be read using the index point labeled “H” on the

lefthand scale and the scale. The Richardson-Jeffes diagrams, although very convenient, cannot be used to obtain more precise values of dissociation pressures. In such cases, rather detailed thermodynamic calculations, described at the beginning of this section, should be performed. Diagrams similar to Fig. 3 are available for formation of sulfides, nitrides, carbides, and halides (Ref 8).

References cited in this section 6. F. Richardson and J. Jeffes, J. Iron Steel Inst., Vol 171, 1952, p 167 7. N. Birks and G.H. Meier, Introduction to High Temperature Oxidation of Metals, Edward Arnold, 1983 8. G.Y. Lai, High-Temperature Corrosion of Engineering Materials, ASM International, 1990


Phase Sequence in the Multilayered Scale Thermodynamic data can be used to predict the local equilibria in the sequence of phases for multilayered oxide scales. In general, if a metal can form several oxides, the sequence of oxides in the scale growing on this metal can be predicted, with the most oxygen-deficient oxide contacting the metal and the most oxygen-rich oxide next to the gas phase. However, not all oxides are stable at a given temperature and so may not be present in the scale. This is the reason why thermodynamic analysis should be used to determine the phase sequence in the multilayered scale. As an example, referring to the Fe-O phase diagram (Fig. 2) and the oxide layer sequence in the scale growing on iron, the following chemical reactions should be considered: (Eq 14) (Eq 15) (Eq 16) For the previous chemical reactions, the dissociation pressure of each oxide can be calculated according to the detailed description given in the first part of this article. At a given temperature, the oxide whose oxygen dissociation pressure is the lowest will form on iron, for example, FeO above 570 °C (1060 °F). To determine whether Fe3O4 or Fe2O3 forms on FeO, the following reaction and corresponding thermodynamic equilibria should be considered: (Eq 17) (Eq 18) The FeO remains at equilibrium with the oxide whose dissociation pressure is lower (Fe3O4). As a consequence, during oxidation by oxygen at temperatures higher than 570 °C (1060 °F), the following phase sequence can be expected on the iron substrate: FeO-Fe3O4-Fe2O3 (Ref 9).

Reference cited in this section 9. P. Kofstad, High Temperature Corrosion, Elsevier, 1988, p 8


Kellogg Diagrams To show a stability range in more complicated multioxidant systems, it is convenient to fix the temperature and plot the other variables, such as gas pressure or alloy composition. These diagrams are called isothermal stability diagrams and show the stability range of particular phases of the system. For the situation of one pure metal and several oxidants, the thermodynamic phase stabilities are shown on log-log plots of the two principal gaseous components, as shown in Fig. 4 for an Fe-O-S system at 727 °C (1341 °F). Such diagrams are constructed on the basis of the chemical potentials of all elements and their compounds that are present in the system. These plots assume that the gas phase is in internal equilibrium with respect to every gaseous species. To plot such a diagram, the equilibria between particular phases of a system should be considered by drawing boundaries, which represent equilibrium between specific oxides or sulfides.

Fig. 4 The thermodynamic phase stabilities in the Fe-O-S system at 727 °C (1341 °F). s, solid; l, liquid For example, the boundary between Fe(s) and FeO(s) phase (Fig. 4) represents the following equilibrium: (Eq 19)

The dissociation pressure of FeO equals 2.3 × 10-22 atm at 727 °C (1341 °F). Next, the boundary between the appropriate oxide (FeO) and sulfide phase (FeS) is constructed from the following equilibrium: (Eq 20)


Determination of Partial Pressures of Gas Mixtures Because the corrosive action of oxygen, hydrogen, or sulfur and their compounds is very common in many essential branches of modern industry, Kellogg diagrams often present auxiliary scales for and so on. However, the composition of a given gas mixture (e.g., SO2 and O2) is a function of temperature. For example, if an iron sample is placed into a container of a SO2-O2 gas mixture with defined composition at room temperature, and then the container is heated to a higher temperature, the values of initial partial pressures of SO2 and O2 cannot be used to determine the presence of oxidation products on iron. If a composition of a gas mixture is known at a given temperature, it is possible to calculate the partial pressures of particular gases at any temperature. To determine the composition of a gas mixture containing S2 and O2 at high temperature, the following equilibria should be considered: (Eq 21) (Eq 22) (Eq 23) (Eq 24) For the previous chemical reactions, the corresponding oxygen partial pressures can be calculated: (Eq 25)

(Eq 26)

(Eq 27)

(Eq 28) Additionally, a mass balance for oxygen and sulfur should be considered:

(Eq 29) (Eq 30) where NS and

denote total amounts of sulfur and oxygen presented in the system (in moles), and and denote the number of moles of a particular gas present at equilibrium in the system at specified temperature. The partial pressure of any gaseous species, pi, can be written as: (Eq 31) where ni is a number of moles of the i-gaseous species, m is the total number of species in the gas mixture, and ptot denotes the total gas pressure (total gas pressure is usually known). Equations 25, 26, 27, 28, 29, 30, 31 can be solved numerically, and the partial pressures of particular gases in the gas mixture can be determined.


Partial Pressures of Volatile Oxidation Products The oxidation reaction product of a given metal may be volatile, liquid, or solid. Even the solid or liquid product could have a high total vapor pressure involving several species. If the oxidation product is volatile, it is useful to determine its partial vapor pressure as a function of oxygen pressure at a given temperature. In the chromium-oxygen system, three volatile oxides can be formed—CrO(g), CrO 2(g), CrO3(g)—in which solid chromium, Cr(s), can also sublime to form gaseous chromium, Cr(g). The only stable oxide in the chromiumoxygen system is Cr2O3(s). The partial pressures of particular volatile species in the chromium-oxygen system for oxygen pressure less than the dissociation pressure of Cr2O3 can be determined from the following equilibria: (Eq 32) (Eq 33) (Eq 34) (Eq 35) In the case of oxygen partial pressure higher than the dissociation pressure of Cr2O3, the partial pressures of the volatile oxides and gaseous chromium can be determined from the appropriate chemical reaction: (Eq 36) (Eq 37)

(Eq 38) (Eq 39) The results of such calculations are presented in Fig. 5. From a practical point of view, it is very important that the partial pressures of volatile CrO2 and CrO3 increase with increasing oxygen pressure and temperature. At temperatures higher than 1000 °C (1830 °F), the weight loss of pure chromium- or chromia (Cr2O3)-forming alloys in air, due to the loss of the volatile CrO2 and CrO3, can be significant. The Cr2O3 layer loses its good protective properties. The reaction kinetics are described by a scaling evaporation mechanism provided in Ref 10.

Fig. 5 The diagram of partial pressures of volatile species in the chromium-oxygen system as a function of oxygen pressure at 727 °C (1341 °F). s, solid; g, gas Theoretical considerations presented in this chapter show that thermodynamics has a fundamental significance in studying the processes occurring in the high-temperature corrosion of metallic materials. Rather simple calculations made on the basis of thermodynamic data determine conditions of temperature, composition, and pressure of reacting gas mixture in which particular reaction products can be formed. As a consequence, thermodynamic considerations may help to optimize the composition of reacting atmosphere and/or alloy and are often the first step in the development of the corrosion resistant materials.

Reference cited in this section 10. C. Tedmon, J. Electrochem. Soc., Vol 113, 1966, p 766


References

1. P. Kofstad, Oxidation Mechanisms for Pure Metals in Single Oxidant Gases, High Temperature Corrosion, R.A. Rapp, Ed., NACE International, 1983, p 123–138 2. Properties of Pure Metals, Properties and Selection: Nonferrous Alloys and Pure Metals, Vol 2, 9th ed., Metals Handbook, American Society for Metals, 1979, p 714–831 3. R.C. Weast, Ed., Physical Constants of Inorganic Compounds, Handbook of Chemistry and Physics, 65th ed., The Chemical Rubber Company, 1984, p B68–B161 4. O. Kubaschewski and B.E. Hopkins, Oxidation of Metals and Alloys, Butterworths, 1962 5. I. Barin, O. Knacke, and O. Kubaschewski, Thermochemical Properties of Inorganic Substances, Springer, 1977 6. F. Richardson and J. Jeffes, J. Iron Steel Inst., Vol 171, 1952, p 167 7. N. Birks and G.H. Meier, Introduction to High Temperature Oxidation of Metals, Edward Arnold, 1983 8. G.Y. Lai, High-Temperature Corrosion of Engineering Materials, ASM International, 1990 9. P. Kofstad, High Temperature Corrosion, Elsevier, 1988, p 8 10. C. Tedmon, J. Electrochem. Soc., Vol 113, 1966, p 766


Selected References • • • • • • • •

N. de Nevers, Physical and Chemical Equilibrium for Chemical Engineers, John Wiley & Sons, Inc., 2002 R.E. Sonntag, C. Borgnakke, and G.J. Van Wylen, Fundamentals of Thermodynamics, John Wiley & Sons, 1998 W. Greiner, Thermodynamics and Statistical Mechanics, Springer, 1995 D.E. Winterbone, Advanced Thermodynamics for Engineers, Arnold, 1997 K. Wark, Jr., Advanced Thermodynamics for Engineers, McGraw-Hill, 1995 D.V. Ragone, Thermodynamics of Materials, John Wiley & Sons, 1995 M. Saad, Thermodynamics: Principles and Practice, Prentice Hall, 1997 P.W. Atkins, The Elements of Physical Chemistry, Oxford University Press, 1996

M. Danielewski, Kinetics of Gaseous Corrosion Processes, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 97–105

Kinetics of Gaseous Corrosion Processes Marek Danielewski, AGH University of Science and Technology, Cracow, Poland

Introduction IN 1923, N.B. Pilling and R.E. Bedworth classified metals into two groups: those that form protective oxide scales and those that do not (Ref 1). They suggested that unprotective scales are formed if the volume of the oxide layer is less than the volume of metal reacted. For example, in the oxidation of aluminum: 2Al + O2 = Al2O3

(Eq 1)

the Pilling-Bedworth molar volume ratio is: (Eq 2) where the volumes are calculated from molecular and atomic weights and the densities of the phases. If the ratio is less than 1, the oxide scales are usually nonprotective. Scales on metals such as magnesium, potassium, sodium, and calcium are porous or crack due to tensile stresses and provide no efficient barrier to penetration of the gas to the metal surface. If the ratio is more than 1, the protective scale may develop and protect the metal or alloy from the gas so that oxidation can proceed only by solid-state diffusion, which is slow even at high temperatures (iron, nickel, cobalt, chromium, silicon, and aluminum and their alloys). If the ratio is over 2, as is the case with tungsten and niobium, during the scale growth large compressive stresses often develop in the oxide that may cause the scale to crack and/or spall off, leaving the metal unprotected. Exceptions to the Pilling-Bedworth classification are numerous. The assumption that metal oxides grow by diffusion of oxygen inward through the oxide layer to the metal is seldom valid. The texture, the direction(s) of scale growth, and the possibility of plastic flow by the oxide or metal were not considered. Nevertheless, historically, Pilling and Bedworth took the first step in understanding of the processes by which metals react with gases. Although there may be exceptions, the volume ratio, as a rough rule-of-thumb, is usually correct. The Pilling- Bedworth volume ratios for many common oxides are listed in Table 2 of the article “Thermodynamics of Gaseous Corrosion” in this Volume.

Reference cited in this section 1. N.B. Pilling and R.E. Bedworth, The Oxidation of Metals at High Temperatures, J. Inst. Met., Vol 29, 1923, p 529–582

M. Danielewski, Kinetics of Gaseous Corrosion Processes, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 97–105 Kinetics of Gaseous Corrosion Processes Marek Danielewski, AGH University of Science and Technology, Cracow, Poland

Defect Structure of Oxides

The basic defect structures of solids were first published by Frenkel (in metals and oxides), Schottky (simple oxides), and Wagner (spinels). Schottky and Wagner started the new discipline, which is now called solid-state chemistry. Point- defect chemistry is the part of it that deals with defect reactions and their equilibria in solids. Such reactions can be written according to Krö ger-Vink notation and must obey the usual rules of electrochemical equations (mass and charge conservancy). The key difference is an additional conservation law, the “rule of the lattice conservancy,” which states that only a stoichiometric number of the cationic and anionic lattice elements can be formed/consumed as a result of a reaction. The meaning of symbols used in the chemistry of point defects can be found in general textbooks or monographs on high-temperature corrosion (Ref 2, 3, 4). Ionic compounds can have appreciable quantities of intrinsic defects (giving rise to ionic conductivity at stoichiometric composition) due to Schottky and/or Frenkel defects (Fig. 1). Schottky defects are combinations of cation vacancies, anion vacancies, and electronic defects: O = VX + VMe

(Eq 3)

where VX and VMe denote the vacancy in anionic sublattice and cationic sublattice, respectively (Fig. 1a). Schottky defects are formed and annihilate at interfaces, and they seldom dominate in oxides that provide protective scales. At high temperatures, ionic defects are ionized and electronic defects are formed in the proper ratio necessary to maintain electrical neutrality: 0 = e- + h· where e- denotes the electron (free or localized) and h· is the electron hole. During oxidation, when the metal is protected by a layer of oxide, electrons migrate from the metal, through the oxide, to adsorbed oxygen at the oxide/gas interface and accelerate the rate of reaction.

Fig. 1 Defects in ionic crystals. (a) Schottky defect. (b) Frenkel defect. Vacancies (VX, VMe) are indicated by open squares. Interstitial ion (Mei) is shown as shaded circle.

Frenkel defects are combinations of cation vacancies, interstitial cations, and electronic defects. The reaction of their formation can be written in the form MeMe = Mei + VMe (Fig. 1b). Frenkel defects are formed/annihilated within the oxide, are the result of “inner sublimation,” and are present in the majority of oxides that protect steels and alloys. Metal cations are generally much smaller than an oxygen anion. Consequently interstitial cations are mobile, showing high diffusivity (mobility). Ionic electrical conductivity is possible in such crystals by the diffusion of interstitial cations and by the diffusion of vacancies. Metallic oxides may have a stoichiometric composition only at specific temperature and pressure conditions. Some oxides always show a certain degree of nonstoichiometry, such as wustite (Fe1-yO). This nonstoichiometry usually implies higher concentrations of defects. The higher-defect concentration leads to the formation of a less-protective scale. A typical example is iron, which can be used at temperatures below 570 °C (1060 °F). At T > 570 °C, the highly defected wustite is formed (Fe1-yO where ~0.05 < y < 0.18), accelerating the oxidation rate by two orders of magnitude. An example of oxides grown to provide protective scales are electronic semiconductors that allow mass transport of ions through the scale layer. They may be categorized as p-type, n-type, and amphoteric semiconductors (Table 1). They always have the other minority defects that contribute to the diffusivity (ionic conductivity). Table 1 Classification of electrical conductors: oxides, sulfides, and nitrides Metallic conductors are in parentheses. Metal-excess semiconductors (n-type) BeO, MgO, CaO, SrO, BaO, BaS, ScN, CeO2, ThO2, UO3, U3O8, TiO2, TiS2, (Ti2S3), TiN, ZrO2, V2O5, (V2S3), VN, Nb2O5, Ta2O5, Cu2O, Cr2S3, MoO3, WO3, WS2, MnO2, Fe2O3, MgFe2O4, NiFe2O4, ZnFe2O4, ZnCo2O4, (CuFeS2), ZnO, CdO, CdS, HgS(red), Al2O3, MgAl2O4, ZnAl2O4, Tl2O3, (In2O3), SiO2, SnO2, PbO2, and at low oxidant pressures Cr2O3, PbS, and MnS Metal-deficit semiconductors (p-type) UO2, (VS), (CrS), Cr2O3, (1250 °C, or 2280 °F), MoO2, FeS2, (OsS2), (IrO2), RuO2, Cr2O3, PbS, and MnS Source: Ref 5 The p-type metal-deficit oxides are nonstoichiometric compounds with cation vacancies being the dominating defects. A typical example is Ni1-yO (Fig. 2), a cation-deficient oxide that provides the additional electrons needed for ionic bonding and electrical neutrality by donating electrons from the 3d subshells of a fraction of the nickel ions (i.e., forms electron holes as electronic defects). The reaction of defects in NiO can be written as O2(g) = OO + + 2h·. In this way, for every cation vacancy present in the oxide, two nickelic ions (Ni3+) will be present. Each Ni3+ has a low-energy positively charged electron hole that electrons from other nickelous ions (Ni2+) can easily move into. The positive or p-type semiconductors carry most of their current by means of these positive holes.

Fig. 2 Ionic arrangement in p-type NiO scale. Cation vacancies are indicated as open squares. The Ni3+ cations (i.e., the electron holes, h) are shaded, the Ni2+ cations are nickel ions in their lattice positions, NiNi. Vacancies are open squares. Cations diffuse through the scale from the Ni/ NiO interface where the nickel atoms enter the oxide: Ni(metallic) + = NiNi + 2e-. They diffuse by cation vacancies (Jc = -JV) to the NiO/gas interface where they react with adsorbed oxygen: NiNi + O2- = + NiNi + OO or traditionally Ni2+ + O2- = NiO. Electrons migrate from the metal surface, by electron holes, to the adsorbed oxygen atoms, which then become chemisorbed oxygen anions: O(adsorbed) + 2e- = O2-. In this way, while cations and electrons move outward through the scale toward the gas, cation vacancies, and electron holes move inward toward the metal. Consequently, as the scale thickens, the cation vacancies may accumulate to form voids at the Ni/ NiO interface and “destroy” an otherwise compact scale. The n-type semiconductor oxides have free electrons as the major charge carriers. They may be either cation excess (Mea+yXb) or anion deficient (MeaXb-y). Beryllium oxide (Be1+yO), a cation-excess oxide, is shown in Fig. 3. Oxygen in the gas adsorbs on the Be1+yO surface and picks up free electrons from the oxide to become chemisorbed O2- ions, which then react with excess metal cations at BeO external interface: + O2- = BeBe + 2+ 2OO (or Be + O = BeO). The cations diffuse interstitially from the underlying beryllium metal. The free electrons migrate from the metal surface where the beryllium enters the oxide and ionizes: Be(metallic) = + 2e-. As with p-type oxides, the cation-excess n-type oxides grow at the oxide/gas interface.

Fig. 3 Ionic arrangement in n-type cation-excess BeO. Interstitial cations Bei are shaded; free electrons are indicated as e-. Another group of n-type semiconductors oxides is anion deficient, as exemplified by zirconium dioxide (ZrO 2y). In this case, although most of the cations contribute four electrons to the ionic bonding, a small fraction of the zirconium cations contribute only two electrons to become the zirconium ion Zr2+. Therefore, to maintain electrical neutrality, an equal number of anion vacancies must be present in the oxide. This arrangement is shown in Fig. 4. The oxide grows at the metal/oxide interface by inward diffusion of O2- through the anion vacancies in the oxide: JO = -JV

(Eq 4)

Fig. 4 Ionic arrangement in n-type anion-deficient ZrO2. Anion vacancies, VZr are indicated as open squares; Zr2+ ions are shaded. Amphoteric Oxides. A number of compounds show nonstoichiometry with either a deficiency of cations or a deficiency of anions. An important example is chromia (Cr2±yO3) (Ref 3) and lead sulfide (Pb1±yS). Both compounds have a minimum in electrical conductivity at the stoichiometric composition and are metal deficient (p-type semiconductors) at the high-oxidant activities and metal excess (n-type) at low pressures. Chromium oxide is a cation-excess oxide at low oxygen pressures and a cation-deficient oxide at high oxygen activities. The mechanism of mass transport in the growing chromia scale depends on the oxygen activity. The oxygen pressure may be low enough to keep the whole chromia layer below the stoichiometric composition (in the n-type range). In such a case, the chromia scale grows in a manner already discussed for n-type semiconductor oxides. At higher oxygen activities the mechanism of diffusion is more complex. Somewhere within the scale there is a transition layer (near-stoichiometric chromia) dividing the oxide into the two “parts,” the n-type semiconductor (from the chromium/chromia interface to near-stoichiometric chromia) and the p-type semiconductor (from the near-stoichiometric chromia to the chromia/gas interface). The mechanism of diffusion and reactions at interfaces depends on the type of oxide that stays in contact with this interface. Oxygen in the gas adsorbs on the chromia surface and picks up free electrons from the chromia to become a chemisorbed O2- ion: O(adsorbed) + 2e- = O2-. At low oxygen partial pressures, this oxygen ion reacts with chromium interstitial ions that are diffusing interstitially from the chromium metal: + 3O2- = 2CrCr + 3OO (or 2Cr3+ + 3O2- = Cr2O3). The free electrons come from the metal surface as the chromium enters chromia and ionizes (Cr(metallic) = ) + 3e-). They can travel through vacant high-energy levels, that is, diffuse by the counterflow of the electron holes: Je = -Jh. As with p-type oxides, the cation-excess n-type chromia grows at the oxide/gas interface as cations diffuse outward through the scale. At high oxygen pressures, chromia in contact + with gaseous oxygen is a p-type oxide. Thus, the reaction at the external interface is: 2CrCr + 3O2- = 3+ 22CrCr + 3OO (or 2Cr + 3O = Cr2O3). Somewhere within the growing chromia there is an intermediate layer where the properties of this oxide change from n-type to p-type. Consequently, outward flow of the interstitial cations converts to the inward vacancy flow (note that this transition of the dominating defects does not affect the direction of the cation flow). Mass transport in amphoteric oxides is not well known; the transition processes may generate stresses and other effects. Grain boundaries and dislocations affect the rate of the transport process.

References cited in this section 2. N. Birks and G.H. Meier, Introduction to High Temperature Oxidation of Metals, Edward Arnold, 1983 3. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, 1988 4. M. Schutze, Corrosion and Environmental Degradation, Vol 1, Materials Science and Technology, A Comprehensive Treatment, R.W. Cahn, P. Haasen, and E.J. Kramer, Ed., Wiley-VCH, 1993 5. O. Kubashewski and B.E. Hopkins, Oxidation of Metals and Alloys, 2nd ed., Butterworths, 1962


Solid-State Diffusion Diffusion processes in solids play a key role in the oxidation of metals. As already stated, the oxidation reaction may be the result of the outward diffusion of metal ions from the metal surface through the oxide layer to the adsorbed oxygen anions at the oxide/gas interface (e.g., p-type NiO), the result of the diffusion of anions inward through the oxide to the metal (e.g., p- type ZrO2) or the result of both processes. The diffusion of atomic oxygen into the metal from the oxide (e.g., into titanium) can also be involved. Within an alloy, interdiffusion plays a key role, for example, the outward diffusion of the reacting metal atoms during their selective oxidation forces the simultaneous inward diffusion of all other alloy elements. Diffusion Mechanisms. Atoms or ions move (diffuse) through solids by many mechanisms. The most common is the vacancy mechanism of diffusion. An atom or ion oscillating (“sitting”) on a regular lattice site can move to a vacant site nearby (diffuse by jumping) (Fig. 5a). In many metal systems, vacancies are the dominating defects, while ionic oxides often show a more complex defect structure. They contain Schottky and Frenkel defects that also involve vacancies and electronic defects. For metal atoms, this is relatively easy because the jump distances are short. For ionic crystals, the jump distances are much longer because cation sites are surrounded by anion sites, and vice versa.

Fig. 5 Diffusion mechanisms. (a) Vacancy diffusion. (b) Interstitial diffusion. (c) Interstitial diffusion with displacement Small interstitial atoms diffuse readily from one interstitial position to another. In ionic oxides, the cations may diffuse interstitially (Fig. 5b), but anions are usually not small enough to do so. In ionic crystals, an interstitial ion often moves into a regular lattice site, knocking another ion into a different interstitial position or to the next

lattice site. This chain effect may extend for several atomic spacings in a line or in different direction. Figure 5(c) shows such interstitial diffusion with displacement. Fick's Law. In 1855, Fick formulated his two laws of diffusion for the simplest sort of diffusion system: a binary system at constant temperature and pressure, with net movement of atoms in only one direction. This is the basic situation for diffusion through an oxide growing on a pure metal. Fick's first law states that the rate of mass transfer is proportional to the concentration gradient of the diffusing element: Ji = -Di Ci

(Eq 5)

where Ji is the flux of the i-th specie (i.e., net mass transported per second through a unit surface area perpendicular to the direction of flow), Ci is the concentration gradient (in one-dimensional, planar systems it becomes:) Ci = i(∂Ci/ ∂x), and Di is the proper diffusion coefficient, cm2/s. The term “proper” means the necessity of the careful examination of the process (experiment) and model of mass transport before using diffusivity data. To model the oxidation processes (i.e., the reactive diffusion), the chemical diffusion coefficients and/or self-diffusion coefficients are useful. To analyze the interdiffusion, the intrinsic and/or self-diffusion coefficients must be known. The ideal Fickian diffusion (Eq 5) is a tracer diffusion, such as an iron isotope in iron. Diffusivity (D) is proportional to effective jump frequency (number of jumps per second when an atom changes position, f) and the square of the jump distance (λ) D = fλ2. More precisely, it depends on the type of diffusing atoms, the chemical bonding, the crystallographic structure of the alloy or oxide, the temperature, and many other factors. When gas oxidation is analyzed in solids, the diffusivities change many orders of magnitude, D (10-18 to 10-6, cm2/ s). Other flux formula were proposed by Nernst- Planck, Onsager, Darken, and others. The Nernst-Planck flux formula is common in electrochemistry and was used by Wagner to analyze the mass transport during the oxidation of the pure metal: Ji = -BiCiF

(Eq 6)

where Bi is the mobility and F is the local force acting on the ith element. The Darken flux formula is common in interdiffusion studies. It was used by Wagner to analyze the reaction of the selective oxidation of the alloy: (Eq 7) where denotes the diffusion flux (given, e.g., by Eq 5 or 6) and v is the drift or convection velocity. Conservation of Mass. The governing law of diffusion, and the mass transport in general, is the law of mass conservation. Its general form is accepted in all instances of mass transport: (Eq 8) where the terms on the right-hand side are the local divergence of the flux of the ith element; Ai is the local source/sink of mass, for example, as a result of chemical reactions. This equation reads: the local change of concentration is the effect of the difference between the local inflow and outflow of mass and a result of chemical reactions. The chemical reactions, the last term, can usually be neglected in solids. In most cases, planar (one-dimensional) systems are analyzed. Thus, the law of mass conservation (or the continuity equation) is: (Eq 9) When the mass transport is due to the diffusion only, Fick's first law is the proper flux formula, and the diffusivity does not depend on concentration, then Eq 8 reduces to the Fick's 1855 formula. The equation of mass conservation (Eq 9) reduces to Fick's second law: (Eq 10) As an example of the Fickian diffusion, the diffusion of oxygen into a metal can be analyzed. When internal oxidation does not occur, the concentration of oxygen changes with time according to the Fick's second law (Eq

10). Solution of the problem depends on the initial and the boundary conditions. When oxygen atoms diffuse inwardly from the surface of an infinite plate, with a constant diffusivity D and at the constant interfacial concentration C′, the solution has a relatively simple form. The concentration of oxygen is a function of distance from the metal surface and of the time [C = C(t, x)]. It is expressed by: (Eq 11) where C0 is a constant, the initial (t = 0) concentration of the oxygen in the plate. The error function, erf, is tabulated in books on diffusion and probability. Figure 6 shows the oxygen concentration as a function of distance from external interface for an arbitrary time (t > 0). During the whole diffusion process, C0 and C′ are constant. One may now ask how fast an arbitrary concentration of oxygen (C″ = constant < C′) changes its “position,” that is, how fast the oxygen penetrates the plate. For the fixed value of oxygen concentration, C″, Eq 11 reduces to: (Eq 12) This means the penetration rate is a parabolic function of time.

Fig. 6 Non-steady-state diffusion. Oxygen distribution during its diffusion into the semi-infinite plate. CM, concentration at metal/oxide interface; C0, initial concentration The Diffusion Coefficient. Diffusivity is proportional to the defect concentration. All defects in a given sublattice contribute to mass transport; for example, the self-diffusion coefficient of metal, , in an oxide is given by: (Eq 13) where denotes the self-diffusion coefficient of the i-th defect and Ni is the molar ratio of defects in Me sublattice. The diffusion coefficient may also depend on crystal orientation (in hexagonal crystals). It implies anisotropic transport properties, and in such a case diffusivity cannot be expressed as a scalar quantity. For oxides that grow epitaxially and/or have a preferred orientation, the diffusivity can differ by orders of magnitude from that of a random polycrystalline oxide. Temperature has a major effect on the diffusion coefficient. In the temperature range where a single mechanism of diffusion dominates and where this mechanism is a thermally activated process, the well-known Arrhenius relation holds. According to the Arrhenius equation, diffusivity increases exponentially with temperature: (Eq 14) where D0 is a constant that depends on frequency of effective jumps, Qa denotes activation energy for diffusion, R is the gas constant, and T the absolute temperature.

Activation energies for interstitial diffusion are lower than those for vacancy diffusion. The activation energy has a major impact on the temperature dependence of a diffusion process. High values of Qa mean that the diffusion proceeds much more rapidly at high temperatures, but might be much slower at low temperatures. If the diffusivity D is plotted on a natural logarithm scale as a function of 1/T, the slope of the resulting straight line is -Qa/R. If the graph shows two intersecting lines, it indicates that the diffusion mechanism that dominates at low temperatures—for example, the grain-boundary diffusion—differs from the one operating at high temperatures (the volume diffusion). Typical values of D0 and Qa for diffusion in oxides are listed in Table 2, (Ref 6). Table 2 Selected diffusion data in metal oxides Metal oxide

Temperature

Copper in Cu2O

°C 800–1050

Nickel in NiO

740–1400

Oxygen in Fe2O3 Iron in Fe3O4

1150– 1250 750–1000

Iron in Fe0.92O

690–1010

0.12

Activation energy for diffusion (Qa) kJ/mol BTU/mol 151.0 143

0.017

234

222

1011

610

578

5.2

230

218

0.014

126.4

120

Frequency factor (D0), cm2/s °F 1470– 1920 1365– 2550 2100– 2280 1380– 1830 1275– 1850 1830– 2460 840–1110 2550– 2910

1000– 4000 420 398 Chromium in 1350 Cr2O3 450–600 2.6 × 10-5 124 118 Oxygen in UO2 1400– 0.25 330 313 Magnesium in 1600 MgO Source: Ref 6 Effect of Impurities. All oxides contain certain substitutional cations (impurities) from the oxidized alloy (present before the oxidation) and/or from the gas phase during oxidation. Although the solubility limit for foreign ions is low, they can have a great effect on the oxide transport properties, and consequently on the oxidation rate. In p-type oxides, such as NiO, the substitutional cations have a valence greater than the Ni2+ ions that are replaced, increasing the concentration of cation vacancies. Two aluminum ions (Al3+) replacing two Ni2+ ions in the cationic sublattice supply two extra free electrons. Consequently, to maintain the local electrical neutrality condition, an additional cation vacancy must be formed. The increase of concentration of cation vacancies in NiO increases the nickel diffusivity in this oxide. On the other side, substitutional cations with a valence lower than +2 reduce diffusion in NiO by reducing the number of cation vacancies. Divalent cations have little effect on diffusion when substituting for Ni2+ ions. For n-type oxides, the effect is reversed; if Al3+ substitutes some titanium ions (Ti4+) in titanium dioxide (TiO2), more anion vacancies will be formed in the oxide. Diffusivity of oxygen then increases because of the increase in anion vacancy concentration. Higher- valence impurity ions would decrease oxygen diffusion in n-type oxides. The impurity effect is especially important for diffusion at low temperatures at which the native defect concentration is low, and the activation energy is associated with the movement of ions only. At high temperatures, the activation energy increases because it involves formation of defects as well as their motion. Fast Diffusion Paths. The activation energies for diffusion along line and surface defects in solids are much lower than those for volume diffusion. Dislocations, grain boundaries, porosity networks, and interfaces form Tm, where Tm is the rapid diffusion paths. At low temperatures, below the Tammann temperature (TT melting temperature), volume diffusion is virtually stopped in oxides. In metals, diffusion along dislocations is more important than volume diffusion below about one-half of the absolute melting point, Tm. Above the Tammann temperature, the volume diffusion predominates in both metals and oxides.

Reference cited in this section 6. T. Rosenqvist, Phase Equilibria in Pyrometallurgy of Sulfide Ores, Metall. Trans. B, Vol 9b, 1978, p 337–351


Oxide Texture Amorphous Oxides. In the very early stages of oxidation and especially at low and intermediate temperatures (T < TT), some oxides appear to grow with an amorphous structure. In general, the oxides in which the molar ratio of oxygen to metal is higher than one form glasses. They contain more oxygen than metal in their formulas so that oxygen triangles or tetrahedra are formed around each of the metal ions. The random network ring structures that result allow large anions or molecular oxygen to move through them more readily than the smaller cations do. Amorphous oxides tend to crystallize as they age. Examples are silicon dioxide (SiO2), tantalum pentoxide (Ta2O5), and niobium pentoxide (Nb2O5). In contrast, oxides with M2O and MO formulas have structures in which the small cations can move readily. They are apparently always crystalline. Examples are NiO, cuprous oxide (Cu2O), and zinc oxide (ZnO). In general the amorphous and nanostructure form of oxides increases mobility and is desirable in coatings and alloys that form protective scale as a result of selective oxidation. In such a case, the stability of the oxide layer depends on continuous supply of oxidized metal; that is, it depends on the interdiffusion rate within an alloy. Epitaxy. As a crystalline oxide grows on a metal surface, it often aligns its crystal structure to be compatible with the structure of the metal substrate. This epitaxial growth results in the best fit between the two different crystal structures. For example, either (111) or (001) planes of Cu2O grow parallel to the Cu (001) plane with the •110• directions of Cu2O parallel to the •110• of copper (Ref 7). Stress develops in an epitaxial oxide layer as it grows because of the slight misfit between the oxide and metal crystals. Such stress is likely to produce dislocation arrays within the oxide that eventually become fast diffusion paths and accelerate mass transport through the film. A mosaic structure may develop in the oxide because of the growth stresses. The mosaic structure consists of small crystallites with orientations very slightly tilted or twisted with respect to each other. The boundaries between the crystallites are dislocation arrays that again serve as fast diffusion paths. Stresses in epitaxial layers increase as the films grow thicker until a point is reached when the bulk scale tends to become polycrystalline and epitaxy is gradually lost. Epitaxy may last up to about 50 nm in many cases, but it seldom exceeds 100 nm. Preferred Orientation. Above the Tammann temperature the oxide grain size (diameter) increases with the scale thickness. Crystals that are favorably oriented can grow at the expense of their neighboring grains until the oxide surface consists of a few large grains with similar orientation. Below the Tammann temperature, finegrained scales are formed. The variation in growth rate of different oxide grains produces roughening of the scale surface and is commonly observed (Ref 2, 3, 4).

References cited in this section 2. N. Birks and G.H. Meier, Introduction to High Temperature Oxidation of Metals, Edward Arnold, 1983 3. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, 1988

4. M. Schutze, Corrosion and Environmental Degradation, Vol 1, Materials Science and Technology, A Comprehensive Treatment, R.W. Cahn, P. Haasen, and E.J. Kramer, Ed., Wiley-VCH, 1993 7. K.R. Lawless and A.T. Gwathmey, The Structure of Oxide Films on Different Faces of a Single Crystal of Copper, Acta Metall., Vol 4, 1956, p 153–163


Oxidation Kinetics Linear Oxidation Reaction Rates. If the metal surface is not protected by an oxide barrier layer, then the oxidation rate usually remains constant with time. In such a case, the surface processes and/or reactions are the rate-controlling steps. This situation occurs in many cases: • • • • •

The Pilling-Bedworth ratio is less than one. The oxide is volatile. Oxide forms molten eutectic with the underlying metal. The scale spalls off or cracks due to internal stresses. A porous, unprotective oxide forms on the metals.

The linear oxidation rate is: (Eq 15) where X is the mass (or thickness) of the oxide formed, t is the time of oxidation, and kL is the linear rate constant. Upon integration and when the initial (at t = 0) scale thickness equals 0, the linear oxidation equation is: X = k Lt

(Eq 16)

The oxidation never slows down; after a short time at high temperature, the metal will be completely destroyed. Figure 7 shows the relationship between oxide mass and time for linear oxidation.

Fig. 7 Types of the oxidation kinetics: linear, logarithmic, inverse logarithmic, and parabolic oxidation kinetics

Logarithmic and Inverse Logarithmic Reaction Rates. At low temperatures when only a thin film of oxide is formed (for example, under 100 nm), the oxidation is usually observed to follow either logarithmic or inverse logarithmic kinetics. Transport processes across the film are rate controlling, with the driving force being electric fields across the film. The logarithmic equation is: X = ke log(at + 1)

(Eq 17)

where ke and a are constants. The inverse logarithmic equation is: (Eq 18) where b and ki are constants. Under the difficult experimental conditions involved in making measurements in the thin-film range, it is difficult to distinguish between logarithmic and inverse logarithmic oxidation. Both Eq 17 and 18 have two constants that can be adjusted to fit the data quite well. Metals oxidizing with logarithmic or inverse log kinetics reach a limiting film thickness at which oxidation apparently stops. Figure 7 shows the curves for both logarithmic and inverse logarithmic kinetics. Parabolic Kinetics. When the rate-controlling step in the oxidation process is the diffusion of ions through a compact barrier layer of oxide, with the chemical potential gradient as the driving force, the parabolic rate law is usually observed. As the oxide grows thicker, the diffusion distance increases and the oxidation rate slows down. The rate is inversely proportional to the oxide thickness: (Eq 19) where k′ is the parabolic rate constant. Upon separating the variables and integrating Eq 19 with the initial condition that at time t = 0 the oxide thickness X = 0, the integral form of the parabolic equation results: X2 = k′t

(Eq 20)

Figure 7 shows the parabolic oxidation curve. Other Reaction Rate Equations. A number of other kinetics equations have been fitted to the experimental data, but it is believed that they describe a combination of the mechanisms described previously, rather than any new basic process. A cubic relationship: X3 = k′t

(Eq 21)

has been reported for very long oxidation periods (t > 10,000 h). It can be shown mathematically to be an intermediate stage between logarithmic and parabolic kinetics. Initial Oxidation Processes: Adsorption and Nucleation. To begin oxidation, oxygen gas is chemisorbed on the metal surface until a complete two-dimensional oxygen layer is formed. Some atomic oxygen also dissolves into the metal at the same time. After the monolayer forms, discrete nuclei of three-dimensional oxide appear on the surface and begin expanding laterally at an ever-increasing rate. The nuclei may originate at structural defects, such as grain boundaries, impurity particles, and dislocations. The concentration of nuclei depends primarily on the crystal orientation of the metal, with more nuclei forming at high pressures and low temperatures. These oxide islands grow outward rapidly by surface diffusion of adsorbed oxygen until a complete film three or four monolayers thick covers the metal. The oxidation rate then drops abruptly. If chemisorption were still the rate-controlling (slow) step in oxidation after the thin film is completed, a logarithmic rate law should be observed. The logarithmic rate law is the result of a strong electric field across the film that affects the oxidation rate.


Thin-Film Mechanisms Many theories have been proposed to explain the oxidation mechanism at low temperatures or in the early stages of high-temperature oxidation where logarithmic kinetics are commonly observed. None of the theories are completely accepted yet, and perhaps none is completely correct, but they have common threads of agreement that indicate reasonably well what is happening. Some of the most important theories are briefly described in this Section. The Cabrera-Mott theory, probably the best established theory of thin-film oxidation, applies to films up to about 10 nm thick (Ref 8). It proposes that electrons from the metal easily pass through the thin film by tunneling to reach adsorbed oxygen at the oxide/gas surface and form oxygen anions. A potential of approximately 1 V is set up between the external oxide surface and the metal. For a film 1 nm thick, the field strength would be 107 V/cm, powerful enough to pull cations from the metal and through the film. The ratecontrolling step is the transfer of cations (or anions) into the oxide or the movement of the ions through the oxide. The electric field reduces this barrier. The structure of the oxide determines whether cations or anions migrate through the oxide. As the film grows thicker, the field strength decreases until it has so little effect on the ions that the rate-controlling mechanism changes. Cabrera and Mott developed an inverse logarithmic kinetic equation to describe the mechanism. A logarithmic equation is more commonly observed, but it can be derived from the Cabrera- Mott mechanism if the activation energy for ionic migration is a function of film thickness. Such a situation would exist if the oxide film were initially amorphous and became more crystalline with aging, giving a constant field strength through the film instead of a constant voltage. The Hauffe-Ilschner theory, a modification of Mott's original concept of a space charge developed across the oxide film, proposes that quantum-mechanical tunneling of electrons is the rate-controlling step (Ref 9). After the film thickness reaches about 10 nm, tunneling becomes increasingly difficult, and the observed reaction rate decreases greatly. For film thickness up to perhaps 20 nm, a logarithmic equation results. For films from 20 to 200 nm thick, the inverse logarithmic relationship holds. Potentials across thin films have been measured; a change in sign of the potential is interpreted as a change from electronic transport control to ionic transport control. The Grimley-Trapnell theory (Ref 10). Grimley and Trapnell used the Cabrera-Mott model, but assumed a constant electric field instead of a constant potential. They assumed that the adsorbed oxygen layer would always be complete, even at high temperatures and low pressure. The adsorbed oxygen would take electrons from cations in the oxide, not from the metal, so that a space charge would develop at the MO/O-ads interface (O-ads being the adsorbed oxygen layer) and be independent of the oxide thickness. If the rate-controlling step is diffusion of cations through vacancies, logarithmic kinetics should be observed. If some other process is rate controlling, linear kinetics is most likely. The Uhlig theory, developed by Uhlig and extended by Fromhold, also predicts logarithmic kinetics at temperatures up to 600 K (Ref 11). The rate-controlling step is the thermal emission of electrons from the metal into the oxide (or electron holes from the adsorbed oxygen ions to the oxide) under the combined effects of induced potential and applied field. The field is created by the diffusing ions. Because growth of the film depends on the electronic work function of the metal, the theory explains oxidation rate changes at crystal and magnetic transformations; most other theories do not. The Wagner Theory of Oxidation. During the oxidation process the two reactants are separated by an oxide layer. Thus, it is necessary to postulate that mass transport (volume diffusion in simple cases) occurs through such growing oxide scale. Before presenting Wagner's derivation of the parabolic rate equation for scale growth

on a metal in which diffusion of ions or electrons is rate-controlling step, a simplified treatment is given to emphasize the main features of diffusion-controlled oxidation. One can assume that a pure metal reacts with an oxidant and the growth of a dense, planar and single-phase oxide occurs. Moreover, the following conditions regarding this phase are necessary for which the theory is valid: • • • • •

One type of defect dominates in the oxide. The thermodynamic equilibria are established on both Me/scale and scale/oxidant interfaces. The oxide scale shows small deviation from stoichiometry. The oxidant solubility in the metal, the oxide evaporation, and any other processes at interfaces are negligible. The scale is thick compared with distances over which space charge effects occur (electrical double layer).

From the assumption of low nonstoichiometry it follows that the metal and oxidant concentrations in the growing scale do not depend on position and are constant. Consequently, when the metal (e.g., interstitial cations Mei) diffusion dominates, Fick's second law (Eq 10) reduces to: (Eq 22) from which it follows that the flux of metal does not depend on position in the scale and depends on time only. The rate of growth of the oxide scale, dX/dt, is then proportional to the flux of metal ions JMe by Fick's first law (Eq 5): (Eq 23) When the diffusion coefficient does not depend on defect concentration, then from Eq 22 and 23 it follows that the gradient of metal concentration in the growing scale is constant. It allows further simplification of the flux formulas: (Eq 24) where both flux and the scale thickness, X, are unknown. The last necessary equation relates the flux of metal at the external interface to the rate of scale formation. It is the mass balance at the scale/oxidant interface and is often called the Stefan condition: (Eq 25) On combining Eq 24 and 25, the differential form of the parabolic rate law follows: (Eq 26) where the parabolic rate constant (term in the brackets) depends on diffusivity of defects and their concentrations. Integrating with the limit that at time t = 0 the oxide thickness X = 0, the integral form of the parabolic equation is: X2 = k′t

(Eq 27)

The Wagner Derivation. Figure 8 gives the reaction scheme for which the theory is valid. Two additional assumptions form the foundation of Wagner's theory of metal oxidation: (1) the migration (diffusion) of ions and electrons across the scale is the rate-controlling process, and (2) it is postulated that thermodynamic equilibrium is established locally through the growing scale. From the first postulate it follows that the more general Nernst-Planck flux formula must be used (instead of Fick's first law), which states that the flow of ions and electrons through a growing oxide layer depends on their concentration, mobility, and the driving force for migration (diffusion):

Ji = CiBiF

(Eq 28)

where Bi denotes the mobility of diffusing specie i and F is the sum of all forces acting on it.

Fig. 8 Diagram of the scale formation according to the Wagner's model A charged ion (valence zi) moving through the oxide is acted upon by two forces: the chemical force (chemical potential gradient, ∂μi/∂x) and an electric potential gradient, ∂φ/∂x. The flux is then: (Eq 29) where Na is Avogadro's number, φ an electrical potential, and F is Faraday's constant. The mobility and selfdiffusion coefficients are related by the Nernst-Einstein relation: Di = BikT. Thus, all of the fluxes in Eq 29 can be written as: (Eq 30)

(Eq 31)

(Eq 32) where R is the gas constant, Nak = R. These flux expressions introduce a new unknown variable. Consequently, an additional equation is necessary. As the oxide forms, the flow of ions through the scale must be balanced by the flow of electrons to maintain a local charge balance: zcJc + zaJa - Je = 0

(Eq 33)

where the subscripts a, c, and e refer to anions, cations, and electrons, respectively. Substituting Eq 30, Eq 31, and 32 into 33 yields: (Eq 34) The postulate of the local thermodynamical equilibrium allows the assumption that everywhere through the oxide, the equilibria: (Eq 35) (Eq 36) are established during the whole process and it follows that: μMe = μc + zcμe

(Eq 37)

μX = μa - |za|μe

(Eq 38)

Moreover, for such isothermal and isobaric process the Gibbs-Duhem relation holds: NMedμMe + NXdμX = 0

(Eq 39)

Upon taking into account that oxides have near- stoichiometric composition, the ratio of molar concentrations may be expressed as: (Eq 40) and Eq 39 becomes |za|dμMe + zcdμX = 0

(Eq 41)

From Eq 37, Eq 38, Eq 39, the electrical potential gradient can be expressed by:

(Eq 42)

Eq 42 reduces to a simple relation: (Eq 43) Introducing Eq 43 into the flux formulas (Eq 30 and 31) and taking into account relations Eq 37 and 38: (Eq 44) (Eq 45) The scale growth is a result of cation and anion fluxes, dX = dXc + dXa. Thus, from the mass balances at both interfaces (Eq 35):

(Eq 46) Introducing Eq 44 and 45, and the Gibbs-Duhem relation (Eq 41), one gets: (Eq 47) The defect diffusion coefficient and self-diffusion coefficients are related through: (Eq 48) Thus, Eq 47 becomes: (Eq 49) The arguments (compare Eq 22, Eq 23, Eq 24, Eq 25, Eq 26 and Eq 49) are valid in the Wagner model. Thus, one can express the right-hand side of Eq 49 using average values. Upon integrating Eq 49 over the entire scale thickness X, from the inner surface (metal/oxide interface), at which the metal chemical potential is at the outer (oxide/gas interface) Eq 49 becomes:

, to

(Eq 50) where (Eq 51) is called the parabolic rate constant. Substituting the Gibbs-Duhem equation: (Eq 52) the result is: (Eq 53) The good agreement between parabolic rate constants calculated from conductivity or diffusivity data and the rate constants measured in oxidation experiments indicates that Wagner's assumptions are generally valid (Ref 2, 3). If the Wagner's principal assumptions about the scale are valid: • • • • •

The oxide scale is completely compact and adherent. The migration of ions through the scale is the rate-controlling process. Thermodynamic equilibrium exists at both the metal/oxide and oxide/gas interfaces. Thermodynamic equilibrium exists locally throughout the scale. The oxide deviates only slightly from stoichiometry.

Then, the Wagner model allows one to understand or use the model as a starting point for more complex reactions, such as the oxidation of alloys. Effects of Temperature and Pressure. When kinetics of the oxidation are controlled by diffusion in the growing scale, the parabolic oxidation rate constant (Eq 19) increases exponentially with temperature, following the Arrhenius equation: (Eq 54) where k0 is a constant dependent on the oxide composition and the gas pressure. According to Wagner's theory, for cation-deficient or cation- excess oxides where Dc » Da, the activation energy Qa for oxide growth is the

same as the activation energy for diffusion of cations in the oxide. For anion-deficient oxides, such as ZrO 2, where Da » Dc, the activation energy for oxide growth is the same as that for anion diffusion, verifying that ionic diffusion is the rate-controlling process. To calculate the effect of pressure on the oxidation rate, Eq 53 can be used. For a cation- deficient p-type oxide, such as Ni1-yO, growth occurs at the oxide/gas interface where: O2(g) + Ni = NiO Using the Kröger-Vink notation, the formation of the nondefect oxide (NiO = NiNi + Oo) is given by O2(g) + Ni = Oo + NiNi where NiNi and Oo indicate that Ni and O occupy the regular positions in the cationic and ionic sublattice. The formation of defects in this notation can be written as: (Eq 55) where symbol stands for a doubly charged cation vacancy, and h· represents an electron hole. Thermodynamic equilibrium is established at the Ni/O 2(g) interface so that at any time the equilibrium constant K for the reaction of the oxide formation (Eq 55) is: (Eq 56) where

is the oxygen partial pressure and the brackets indicate the concentration of the species enclosed; for

example, denotes the concentration of cation vacancies. The oxygen concentration at both interfaces is constant. Thus, the local thermodynamic equilibrium conditions imply that locally there must be two electron holes for every cation vacancy, or [h·] = 2 [ ], then: (Eq 57) The nickel diffusivity is proportional to the vacancy concentration (Eq 13), so that: (Eq 58) where D0 denotes here the cation diffusivity for oxygen pressure equal one. It may be shown that for an oxide with simple defect structure (where only one type of defect dominates in a cationic or anionic sublattice) Eq 58 takes the form: (Eq 59) where, theoretically, both the sign and n (generally a number between 2 and 8) depend on the type of dominating defect. Substituting Eq 58 into Eq 53, the oxidation rate is: (Eq 60) which upon integration becomes: (Eq 61) In most cases, the ambient oxygen pressure so that:

is much greater than the nickel oxide dissociation pressure

(Eq 62) Equation 62 explains both the Arrhenius-type temperature dependence and the pressure dependence of the oxidation rate. For an oxide with a simple defect structure and where the concentration of defects at the Me/O2(g) interface is higher than at the oxide/Me interface, Eq 62 has the form:

(Eq 63) For oxides with a simple defect structure and when the concentration of defects at the Me/ O2(g) interface is lower than at oxide/Me interface—for example, Cu2O—the oxidation rate is practically independent of the ambient oxygen pressure.

References cited in this section 2. N. Birks and G.H. Meier, Introduction to High Temperature Oxidation of Metals, Edward Arnold, 1983 3. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, 1988 8. N. Cabrera and N.F. Mott, Theory of Oxidation of Metals, Rep. Prog. Phys., Vol 12, 1948–1949, p 163– 184 9. K. Hauffe and B. Ilschner, Defective-Array States and Transport Processes in Ionic Crystals, Z. Electrochem., Vol 58, 1954, p 467–477 10. T.B. Grimley and B.M.W. Trapnell, The Gas/Oxide Interface and the Oxidation of Metals, Proc. R. Soc. (London) A, Vol A234, 1956, p 405–418 11. H.H. Uhlig, Initial Oxidation Rate of Metals and the Logarithmic Equation, Acta Metall., Vol 4, 1956, p 541–554


Acknowledgment Portions of this article have been adapted from S.A. Bradford, Fundamentals of Corrosion in Gases, Corrosion, Vol 13, ASM Handbook (formerly 9th ed. Metals Handbook), ASM International, 1987, p 61–76.


References

1. N.B. Pilling and R.E. Bedworth, The Oxidation of Metals at High Temperatures, J. Inst. Met., Vol 29, 1923, p 529–582 2. N. Birks and G.H. Meier, Introduction to High Temperature Oxidation of Metals, Edward Arnold, 1983 3. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, 1988 4. M. Schutze, Corrosion and Environmental Degradation, Vol 1, Materials Science and Technology, A Comprehensive Treatment, R.W. Cahn, P. Haasen, and E.J. Kramer, Ed., Wiley-VCH, 1993 5. O. Kubashewski and B.E. Hopkins, Oxidation of Metals and Alloys, 2nd ed., Butterworths, 1962 6. T. Rosenqvist, Phase Equilibria in Pyrometallurgy of Sulfide Ores, Metall. Trans. B, Vol 9b, 1978, p 337–351 7. K.R. Lawless and A.T. Gwathmey, The Structure of Oxide Films on Different Faces of a Single Crystal of Copper, Acta Metall., Vol 4, 1956, p 153–163 8. N. Cabrera and N.F. Mott, Theory of Oxidation of Metals, Rep. Prog. Phys., Vol 12, 1948–1949, p 163– 184 9. K. Hauffe and B. Ilschner, Defective-Array States and Transport Processes in Ionic Crystals, Z. Electrochem., Vol 58, 1954, p 467–477 10. T.B. Grimley and B.M.W. Trapnell, The Gas/Oxide Interface and the Oxidation of Metals, Proc. R. Soc. (London) A, Vol A234, 1956, p 405–418 11. H.H. Uhlig, Initial Oxidation Rate of Metals and the Logarithmic Equation, Acta Metall., Vol 4, 1956, p 541–554 12. C. Wagner, Contributions to the Theory of the Tarnishing Process, Z. Phys. Chem., Vol B21, 1933, p 25–41 13. R.C. Weast, Ed., CRC Handbook of Chemistry and Physics, 62nd ed., CRC Press, 1981 14. R.B. Ross, Ed., Metallic Materials Specification Handbook, 4th ed., Chapman & Hall, London, 1992 15. Properties and Selection: Nonferrous Alloys and Special-Purpose Materials, Vol 2, ASM Handbook, ASM International, 1990 16. A. Nayer, Ed., The Metals Databook, McGraw-Hill, 1997 17. D.R. Lide, Ed., CRC Handbook of Chemistry and Physics, 79th ed., CRC Press, 1998


Selected References • • • • • •

N. Birks and G.H. Meier, Introduction to High Temperature Oxidation of Metals, Edward Arnold, 1983 A.T. Frumhold, Jr., Theory of Metal Oxidation, Vol 1—Fundamentals, North Holland Publishing, 1976 A.S. Khanna, Introduction to High Temperature Oxidation and Corrosion, ASM International, 2002 P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, 1988 H. Schmalzried, Chemical Kinetics of Solids, VCH, 1995 M. Shutze, Corrosion and Environmental Degradation, Vol 1, Materials Science and Technology, A Comprehensive Treatment, R.W. Cahn, P. Haasen, E.J. Kramer, Ed., Wiley-VCH, 1993

M. Danielewski, Gaseous Corrosion Mechanisms, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 106–114

Gaseous Corrosion Mechanisms Revised by Marek Danielewski, AGH University of Science and Technology, Cracow, Poland

Introduction THE CHARACTERISTICS AND BEHAVIOR of scale produced by various types of oxidation are examined in this article. The basic models, concepts, processes and open questions for high-temperature gaseous corrosion are presented.

M. Danielewski, Gaseous Corrosion Mechanisms, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 106–114 Gaseous Corrosion Mechanisms Revised by Marek Danielewski, AGH University of Science and Technology, Cracow, Poland

Properties of Scales Multiple Scale Layers. A pure metal that can be oxidized to more than one valence state can form a series of oxides in the form of separate layers. For example, iron forms highly defected wustite (FeO), magnetite (Fe3O4), and (at high oxygen pressure) hematite (Fe2O3) scale layers. The layers will be arranged with the most metal- rich oxide next to the metal and the most oxygen-rich layer on the outside. At temperatures above 583

°C (1081 °F), wustite is stable, and the sequence of oxides in the scale shows the following pattern: Fe|FeO|Fe3O4|Fe2O3|O2(g). At temperatures below 583 °C (1081 °F), rapidly growing wustite is not stable, and consequently, iron shows satisfactory resistance to oxidation. The corrosion-resistant scale formed below 583 °C (1081 °F) on iron consists of magnetite and hematite: Fe|Fe3O4|Fe2O3|O2(g). A concentration gradient exists within each layer, with higher metal ion concentration closest to the metal. Another example is the copper-oxygen system, which may form the following sequence of oxides: Cu|Cu2O|CuO|O2(g). If the oxygen partial pressure in the gas is below the dissociation pressures of the outer oxygen-rich oxides, then only the thermodynamically stable, inner oxide(s) will form. In general, the lowest valence, inner oxide will usually show the defects in a cationic sublattice only (vacancies in wustite: Fe 1-yO, but interstitial cations in Cr2+yO3 and Cu2+yO). The highest-valence oxide will usually show the vacancies in cationic and/or defects in anionic sublattices. The nonstoichiometry of such oxides (e.g., Fe2O3) is very low and often nonmeasurable. Consequently, the defect structure and transport properties of many technically important oxides are not very well known (e.g., Al2O3). A scale consisting of an inner layer with cations diffusing outward and an outer layer with anions diffusing inward will grow at the oxide-oxide interface. Relative Thickness. When diffusion is the rate-controlling process and no porosity develops in the oxide layers, their relative thickness is proportional to the relative diffusion rates. For compact layers growing by a single diffusion mechanism, the ratio of thicknesses should be related to the ratio of the parabolic rate constants by: (Eq 1) where subscripts 1 and 2 refer to layers 1 and 2. The thickness ratio, consequently, is a constant and does not change with time. Because the ions diffusing through the various layers are likely to be different, and because the crystal structures of the layers are certainly different, the thickness ratio is commonly found to be quite far from unity. One layer is usually much thicker than the others. When diffusion controls the growth of each layer, the entire scale will appear to follow the parabolic oxidation equation, with an effective parabolic rate constant . However, this effective parabolic rate constant does not need to follow the Arrhenius equation (Eq. 54 in the article “Kinetics of Gaseous Corrosion Processes” in this Volume) unless the thickness ratios remain far from unity throughout the temperature range, that is, unless one layer predominates in the analyzed temperature range. If an inner scale predominates, its growth is independent of oxygen pressure, but if the outermost scale is the major part of the scale, the rate constant will vary with oxygen pressure. Paralinear Oxidation. With some metals, the oxidation growth is parabolic at the beginning, but the protective scale gradually changes (as a whole layer or only partially) to a nonprotective layer. If the inner protective layer remains at a constant thickness, then the diffusion through this layer results in a linear rate of oxidation. The outer layer may become nonprotective by sublimation, transformation to a porous layer, fracture, and so on. This type of oxidation behavior that is initially parabolic and gradually transforms to linear is termed paralinear oxidation (Fig. 1).

Fig. 1 Paralinear oxidation. Scale growth (mass) is initially parabolic and becomes linear with time.

Oxide Evaporation. At high temperatures, the evaporation of a protective oxide may limit its protective qualities or remove the oxide entirely, because the evaporation rate increases exponentially with temperature. Platinum alloys and refractory metals in particular tend to have volatile oxides. Suboxides and unusual valences are also often found at high temperatures; chromium, for example, forms only one stable solid oxide, Cr 2O3, but vaporizes as CrO3 at temperatures above 1173 K (1652 °F). Reactive evaporation in the chromium-oxygen system was measured experimentally and explained theoretically (Ref 1). Evaporation may be much more intense in gases containing water or halide vapor if volatile hydroxides (hydrated oxides) or oxyhalides form. Theoretically, at low pressures when evaporating molecules do not return to the surface (do not resublimate), the evaporation rate is directly proportional to the sublimation vapor pressure of the oxide. In practical conditions, the total pressure often exceeds 1 atm, while the low oxygen partial pressure creates conditions for oxide evaporation. In such conditions at gas pressures above 10-3 to 10-4 atm and when the velocity of gas is low (laminar flow), a gaseous stagnant boundary layer slows the escape of the evaporated oxide molecules (stimulates resublimation). The boundary layer becomes thinner at higher gas velocities, leading to higher evaporation losses. Reactive diffusion provides parabolic oxide growth. The rate of the scale growth decreases (Eq. 53 in the article “Kinetics of Gaseous Corrosion Processes” in this Volume), while the evaporation removes material from the oxide layer at a constant rate (time-independent process). The rate of diffusion through the oxide decreases until the two rates finally become equal. The oxide thickness then remains constant, and the metal oxidizes linearly; that is, the metal consumption shows linear time dependence. If more than one oxide layer protects the metal, the higher-valent, outermost oxide usually has the higher vapor pressure and is more volatile. A detailed analysis of evaporation processes and their impact on oxidation is presented in Ref 2.

References cited in this section 1. K.P. Lillerud and P. Kofstad, J. Electrochem. Soc., Vol 127, 1980, p 2397 2. N. Birks and G.H. Meier, Introduction to High Temperature Oxidation of Metals, Edward Arnold, 1983


Stresses in Scales Stress development is inherently coupled with scale growth processes and is often the key factor limiting the protective properties of the scale. During scale growth, recrystallization may occur in either the alloy or the oxide, altering the stresses radically. Changes in temperature usually have a great effect on the stress state for metals in such service. Numerous analyses of practical damage cases have shown that the failure of components is often initiated by stresses resulting from either the oxidation process itself or from operation of the oxidized element. In the past, stress development and relief were often neglected in laboratory investigations; today, the role of stress is well recognized (Ref 2, 3, 4). There are generally three types of stress that have to be considered: • • •

External stresses, σext, resulting from operation Thermally induced stresses, σtherm, due to temperature changes of the system Growth stresses, σox, resulting from the oxidation/corrosion process itself

The sum of all stresses determines mechanical scale failure. If the sum reaches or exceeds a certain critical value, σc, cracking or spalling of the protective oxide follows: σtot = σext + σtherm + σox ≥ σc

(Eq 2)

Stresses can be measured by different experimental techniques and also predicted from model calculations to assess whether scale failure is likely (Ref 4). Growth Stresses. The Pilling-Bedworth ratio, first published in 1925, was considered as a major factor for determining the magnitude of growth stress. The Pilling-Bedworth ratio is the ratio of the oxide volume per mole of the metal in the oxide to the metal volume per mole of metal in the metal or alloy. Numerous observations have shown that each step of the scale growth process can generate stress: •

• • • •

• • •

Sometimes crystalline oxides grow on a metal substrate with an epitaxial relationship. As a result of the different lattice parameters, the stresses that develop often limit the epitaxy to several nanometers of scale. Polycrystalline oxides develop stresses along their grain boundaries as a consequence of growth of grains in the direction perpendicular to or different from the direction of the scale growth. Grain-boundary diffusion, diffusion along oxide grain boundaries, may lead to oxide formation within the growing scale at the boundaries and create compressive stresses. Any second phases or foreign inclusions in a metal may oxidize at a rate different from that of the parent metal and create high stresses within the oxide. Local changes in composition of the growing scale (due to large deviations from stoichiometry) may result in tensile stress; wustite is an important example, varying from Fe0.95O in equilibrium with the metal to as little as Fe0.84O in equilibrium with Fe3O4 at 1370 °C (2500 °F). Interdiffusion in a reacting alloy can generate stresses (i.e., Kirkendall effect) when alloy components have different mobilities. Diffusion of oxygen into the metal from the oxide creates compressive stresses in the metal, such as in the titanium-oxygen system. Surface geometry contributes an additional effect to the growth stresses in an adherent oxide, depending on the surface profile.

Figure 2 shows the four possibilities for changes in stress state in an oxide scale as it grows, assuming that the original growth stresses are compressive (Ref 5).

Fig. 2 The development of geometrically induced growth stresses (horizontal arrows) in oxide scales for four combinations of curved surfaces (convex and concave), and growth by anion (left) and by cation (right) diffusion. R, radius of curvature; M, oxide displacement vector; a, volume fraction of oxide formed at the interface. Details of (a) to (d) can be found in the text. Source: Ref 5

For a convex surface on which oxide grows at the oxide-gas interface by cation diffusion outward through the scale (Fig. 2b), the metal surface (M) will gradually recede, increasing the compressive stresses at the metaloxide interface as long as adhesion is maintained. Figure 2(a) shows oxidation of a convex surface on which the oxide grows at the metal-oxide interface by anion diffusion inward and generates compressive stresses at the metal-oxide interface. For concave surfaces (Fig. 2d), oxide grows at the oxide-gas interface by outward diffusion of cations. As the metal surface recedes, the compressive growth stresses are reduced and may eventually even become tensile if oxidation continues long enough. Figure 2(c) shows growth on a concave surface by anion diffusion inward for reaction at the metal-oxide interface. Very high compressive stresses develop during growth, until they exceed the cohesive strength of the oxide. Two major types of growth stresses are distinguished. The first type is the geometrically induced growth stresses that are due to the surface curvature or components (Fig. 2), and the second type is the intrinsic growth stresses. As can often be seen in oxidation experiments on flat samples, the oxide scales crack at the edges of the specimens, for example, during tungsten oxidation. This type of growth stress has been analyzed (Ref 6). With the help of models, the tangential and radial stresses can be calculated for the ideal case of curved surfaces with a constant radius of curvature. The oxidation increases the strain in the circumferential direction (tangential strain,

) at the rate: (Eq 3)

where Rs denotes the radius of curvature of the surface (concave, Rs < 0; convex, Rs > 0), h is the metal recession, X denotes scale thickness, and RPB is the Pilling-Bedworth ratio. The oxide displacement vector, M, can be calculated from the following expression: M = RPB(1 - a) - (1 - V)

(Eq 4)

where a is the volume fraction of oxide formed at the oxide-gas interface and (1 - a) at the metal-oxide interface, and V and (1 - V) denote, respectively, the volume fraction of the metal consumed as a result of the injection of vacancies into the oxidized metal and its volume fraction consumed during oxide formation at the metal-oxide interface. The scale thickness and metal recession are related: X = RPBh

(Eq 5)

Equation 3 allows one to calculate the tangential stresses, , when linear elastic behavior can be assumed. The magnitude of the maximum radial stresses, , is given by: (Eq 6) The signs of the tangential and radial strains and stresses in the scale and at the metal-oxide interface are given by: (Eq 7) (Eq 8) where “+” and “-” denote tensile and compressive stresses, respectively, and Rs is “+” if convex and “-” if concave. At present, a relatively limited understanding of the mechanisms leading to intrinsic growth stresses has been achieved. The growth stresses were computed for alumina formers (FeCrAl and FeCrAlY alloys) and chromia formers (chromium, NiCr, and FeNiCr stainless steel); oxidation of pure nickel cannot be accurately predicted using model calculations (Ref 4). Mechanical scale failure takes place when a critical stress level, σc, reached. This critical stress level can be converted into a critical strain, εc, by dividing the stress by the oxide Young's modulus, Eox, if it is assumed that elastic behavior dominates in scale failure. Strain values are more easily accessible than stress values, because they can be determined by experiments.

Generally, two modes of failure dominate, depending on the stress distribution in the scale: the failure due to tensile stresses and the failure due to compressive stresses. For tensile failure, a mechanical model has been developed describing the critical strain, εc: (Eq 9) where KIc denotes the fracture toughness of the scale, f is a geometrical factor, and c denotes the radius of the defect (pore, crack, precipitate, etc.). Applying this equation, a researcher computed the critical strains for some common oxides (Ref 4). Transformation Stresses. Preferential oxidation of one component in an alloy may alter the alloy composition to the point that a crystallographic phase transformation occurs. A change in temperature could also cause crystallographic transformation of either the metal or the oxide. The volume change accompanying a transformation creates severe stresses in both the metal and the oxide. Some oxides initially form an amorphous/nanocrystalline structure and gradually crystallize as the film grows thicker. The tensile stress created by volume contraction may partially counteract the compressive growth stresses usually present. Thermal Stresses. A common cause of failure of oxide protective scales is the stress created by cooling from the reaction temperature. The stress generated in the oxide is directly proportional to the difference in coefficients of linear expansion between the oxide and the metal. Coefficients are listed in Table 1 for a few important metal-oxide systems. In most cases, the thermal expansion of the oxide is less than that of the metal; therefore, compressive stress develops in the oxide during cooling. Multilayered scales will develop additional stresses at the oxide-oxide interface. Table 1 Coefficients of linear thermal expansion (CTE) of metals and oxides System Fe/FeO Fe/Fe2O3 Ni/NiO Co/CoO Cr/Cr2O3 Cu/Cu2O Cu/CuO Source: Ref 7

Oxide coefficient 10-6/K 10-6/°F 12.2 6.78 14.9 8.28 17.1 9.50 15.0 8.3 7.3 4.1 4.3 2.4 9.3 5.2

Metal coefficient 10-6/K 10-6/°F 15.3 8.50 15.3 8.50 17.6 9.78 14.0 7.8 9.5 5.3 18.6 10.3 18.6 10.3

Temperature range °C °F 100–900 212–1650 20–900 70–1650 20–1000 70–1830 25–350 75–660 100–1000 212–1830 20–750 70–1380 20–600 70–1110

References cited in this section 2. N. Birks and G.H. Meier, Introduction to High Temperature Oxidation of Metals, Edward Arnold, 1983 3. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, 1988 4. M. Schütze, in Corrosion and Environmental Degradation, Vol 1, Materials Science and Technology, A Comprehensive Treatment, R.W. Cahn, P. Haasen, and E.J. Kramer, Ed., Wiley-VCH, 1993 5. M.I. Manning, Corros. Sci., Vol 21, 1981, p 301 6. W. Christl, A. Rahmel, and M. Schütze, Oxid. Met., Vol 31, 1989, p 1 7. P. Hancock and R.C. Hurst, The Mechanical Properties and Breakdown of Surface Films at Elevated Temperatures, Advances in Corrosion Science and Technology, Vol 4, R.W. Staehle and M.G. Fontona, Ed., Plenum Press, 1974, p 1–84


Stress Relief The oxide may develop porosity as it grows and relieves stress. If temperatures are high, the stress may reach the yield strength of either the metal or the oxide, so that plastic deformation relieves the stress. A brittle oxide may crack. A strong oxide may remain intact until the internal stresses exceed the adhesive forces between metal and oxide, so that the oxide pulls loose from the metal. These ways in which stresses can be relieved are discussed subsequently. Porosity. For oxides that grow by cation diffusion outward through cation vacancies (p-type oxides, such as NiO), vacancies are created at the oxide-gas interface and diffuse inward through the scale as they exchange places with an equal number of outward-diffusing cations. The vacancies are annihilated within the oxide, at the metal-oxide interface, or within the metal, depending on the system. In some oxides, the vacancies collect together within the oxide to form approximately spherical cavities. The preferred sites for cavity formation are along paths of rapid diffusion, such as grain boundaries and dislocation lines. Vacancies that are annihilated at the metal-oxide interface may cause detachment of the scale, because voids that form there reduce adhesion of oxide to metal. If detachment is only partial, the oxidation rate slows down, because the cross- sectional area available for diffusion decreases. Plastic Flow of Oxide. Dislocations are formed in the oxide as it grows epitaxially on the metal because of the crystallographic lattice mismatch between oxide and metal. As the growth continues, these glissile slip dislocations move out into the oxide by the process of glide. Once out in the oxide, they become sessile growth dislocations. Although dislocations are present in the oxide, slip is not an important process in relieving growth stress (Ref 8). Plastic deformation of the oxide occurs only at high temperatures, at which creep mechanisms become operative. The three important creep mechanisms in oxides are grain- boundary sliding, Herring-Nabarro creep, and climb. Grain-boundary sliding allows relative motion along the inherently weak boundaries. HerringNabarro creep allows grain elongation by diffusion of ions away from grain-boundary areas in compression over to boundaries in tension. Within the grains, dislocation climb is controlled by diffusion of the slowermoving ions. The creep rate increases with the amount of porosity in the oxide. Cracking of Oxide. Tensile cracks relieve growth stresses in the oxide scale. As shown in Fig. 2, tensile stresses may eventually develop in an oxide growing on curved surfaces and cause fracture. In the case of anion diffusion inward on convex surfaces, if oxygen diffuses into the metal, the tensile stresses near the oxide-gas interface develop much more quickly. Shear cracks can form in an oxide having high compressive stresses near the metal surface, if the scale cohesion is weak and adhesion to the metal is strong. If the metal-oxide interface is planar, a shear crack initiating at the interface can extend rapidly across the surface. However, if the interface is rough because oxidation has concentrated at the grain boundaries of the metal and keyed the oxide into the metal, rapid crack extension may be prevented and scale adherence may be improved. Periodic cracking of a protective oxide results in the parabolic oxidation being interrupted by a sudden increase in rate when the gas can react directly with the bare metal surface. As oxide begins to cover the metal surface again, parabolic oxidation is resumed. A typical oxidation curve for this repeated process is shown in Fig. 3(a). The time periods between successive parabolic steps are sometimes fairly uniform, because a critical scale thickness is reached that causes cracks to be initiated. The overall oxidation of the metal approaches a slow linear process.

Fig. 3 Cracking of oxide. (a) Periodic cracking of the scale. (b) Breakaway oxidation. (c) Morphology of the single-phase double-layer scale Occasionally, a metal oxidizes parabolically until the scale cracks or spalls off, and from that time on, the oxidation is linear. The oxide, originally protective, completely loses its protective properties. The breakaway oxidation commonly occurs if many cracks form continuously and extend quickly through the oxide. It can also occur for alloys that have had one component selectively oxidized. When the protective scale spalls off, the metal surface is so depleted in the component that the same protective scale cannot reform. Breakaway oxidation (Fig. 3b) leaves bare metal continually exposed, unlike paralinear oxidation (Fig. 1), in which an inner protective scale always remains. Decohesion and Double-Layer Formation. For scale growth by cation diffusion, a protective scale may eventually reach a thickness at which it can no longer deform plastically to conform to the receding metal surface. At this point, decohesion begins at some places along the metal- oxide interface. However, oxidation continues at the oxide-gas interface because cations continue to diffuse outward through the detached scale, driven by the chemical potential gradient across the scale. At the inner scale surface, the cation concentration decreases, thus locally increasing the oxygen chemical potential. The increased chemical potential results in an increased pressure of O2 gas that will form in the space between the scale and the metal. The O2 is then transported to the metal surface by means of surface diffusion and/ or gas phase migration. Because its pressure is greater than the dissociation pressure of oxide in equilibrium with metal, an inner layer of oxide begins to form on the metal. This mechanism is called the dissociation mechanism of the scale growth (Ref 9). The inner oxide layer forms a porous, finegrained structure and has essentially the same composition as the compact outer layer. Initially, it starts forming at the nucleation sites of bare metal. The diffusion of O2 from the outer layer to the inner layer is rate controlling, so that the inner layer grows mainly at its high points and forms a porous scale (Ref 9). Figure 3(c) shows the mechanism of the double-layer scale growth. For the oxidation of pure chromium, it has been found that at temperatures above 1000 °C (1800 °F), the scale cracks periodically when the intrinsic growth stresses exceed a critical value (Ref 1). Several parabolic partial steps exist in the kinetics curve, which are due to cracking after each of these steps and exposure of the bare metal surface under a detached and cracked oxide scale to the surrounding atmosphere. Thus, the fast, initial oxidation period starts again, leading to a new parabolic partial oxidation curve if the temperature is increased to 1200 °C (2200 °F); the intervals for cracking become so short that they can no longer be resolved in the curve, and a fast, linear curve results. This shows that permanent cracking of the scale by intrinsic growth stresses allows no protection, and therefore, the initial surface oxidation steps are always rate determining. The buckling and cracking of oxide scales due to intrinsic growth stresses can be suppressed by the addition of active-elements to the alloy. There are several possible explanations for the active-element effect, which helps to improve adhesion of the scale and which has been proven for a number of different materials. The major mechanisms are changes of the transport properties in the scale; that is, metal cation diffusion in the outward direction can be suppressed by the presence of certain active elements so that only oxygen-inward transport occurs, leading to scale formation solely at the metal interface and reducing the growth stresses. Numerous mechanisms to explain the reactive- element effect were proposed (Ref 10): •

A change in the transport properties of the oxide; for example, suppression of cation diffusion results in a decreased oxidation rate, and the oxygen-inward diffusion becomes the dominating mode of transport

• •

• • •

in the growing scale. As a result of the oxide formation at the metal-oxide interface, the growth stresses are reduced. An increase of the density of the nucleation sites at the metal surface results in finegrained scale, leading to a more plastic scale. Highly reactive elements react with sulfur impurities in the metal substrate, thus impeding the segregation of sulfur to the metal-oxide interface, which would otherwise reduce the adhesion of the scale. The formation of oxide pegs in the metal (at the metal-oxide interface, for example, at the metal grain boundaries) leads to the keying of the oxide to the metal. Reducing the number of physical defects, such as pores in the oxide, decreases the susceptibility of the scale to cracking or detachment. The precipitation of oxides containing reactive elements reduces the fast diffusion along the grain boundaries.

Advanced materials and coatings used at temperatures near 1000 °C (1800 °F) benefit from the effect of activeelement additions. The active element plays the key role in providing long- term oxidation protection of the protective coatings on gas turbine materials. The most common active elements are yttrium, cerium, rhenium, hafnium, and zirconium. Deformation of Metal. Foils and thin-wall tubes are often observed to deform during oxidation, thus relieving the oxide growth stresses. At high temperatures, slip and creep mechanisms can be operative in metals while the temperature is still too low for plastic deformation of the oxide. The deformation processes are facilitated by accumulation of porosity in the metal, which is caused by outward cation diffusion (e.g., accumulation/injection of vacancies) and by the selective oxidation of one component of an alloy (e.g., Kirkendall effect in the alloy near the metal-oxide interface). Catastrophic Oxidation. Although many oxidation failures can be described as catastrophes, the term catastrophic oxidation was originally reserved for the special situation in which a liquid phase is formed in the oxidation process. This can occur either when the metal is exposed to the vapors of a low-melting oxide or during oxidation of an alloy having a component that forms a low-melting oxide or sulfide (see Table 2 in the article “Thermodynamics of Gaseous Corrosion” in this Volume). The mechanisms of catastrophic oxidation vary, but the evidence shows that it occurs at the metal-oxide interface. A liquid phase seems to be essential. The liquid usually forms at the oxide-gas surface and penetrates the scale along grain boundaries or pores to reach the metal. The penetration paths can also serve as paths for rapid diffusion of reacting ions. Once at the metal-oxide interface, the liquid spreads out by capillary action, destroying adherence of the solid scale. In the case of alloys, the liquid can penetrate along the grain boundaries and cause intergranular corrosion. When diffusion and/or flow of the oxidant in the gas phase is the rate-limiting process, then oxidation in the absence of a protective scale proceeds linearly. In a case of fast reactions, the metal and corrosion products heat up from the exothermic reaction, and the oxidation process can be even more rapid. Internal Oxidation/Sulfidation. Internal oxidation (sulfidation) is the term used to describe the formation of fine oxide or sulfide precipitates within an alloy. The oxidant dissolves in the alloy at the metal-oxide interface or at the bare metal surface if the gas pressure is below the dissociation pressure of the metal oxides. Subsequently, it diffuses into the metal and forms the most stable oxide that it can. This is usually the oxide of the most reactive component of the alloy. Internal oxides can form if the reactive element diffuses outward more slowly than the oxygen diffuses inward; otherwise, the surface scale would form. The transport (diffusion) of internally oxidized metal is controlled by the interdiffusion in the alloy. Oxygen transport in an alloy can be treated as Fickian diffusion (Eq 5 and 8 in the article “Kinetics of Gaseous Corrosion Processes” in this Volume). Because the diffusion coefficients of oxidized metal and oxidant vary exponentially with temperature, and because their activation energies are different, it is possible that internal oxidation will occur in an alloy only in a certain temperature range. Because diffusion of oxygen is usually the rate-controlling process in internal oxidation, parabolic behavior is often observed. Based on the assumption that interdiffusion of reactive metal B atoms is negligible and that oxygen has a very low solubility limit in the alloy, the simple quasi- stationary model of the process was formulated. A simple parabolic equation was derived that allows an estimation of the rate of the process (Ref 11):

(Eq 10) where is the mole fraction of oxygen in the alloy at its surface, is the initial mole fraction of reactive metal B in the alloy, Do is the diffusion coefficient of oxygen in the alloy, ν is the ratio of oxygen atoms to metal atoms in the reaction product, and X is the thickness of the internal oxidation zone (i.e., the subscale). The real processes encountered in practice are more complex. In a binary alloy, a mixed (A,B)O oxide may form as the internal oxide. This will occur if AO and BO have considerable mutual solubility so that the free energy of the system is lowered by precipitation of the mixed oxide. For example, the internal oxide that forms in unalloyed steels is (Fe,Mn)O (Ref 12). When a cast 25Cr-20Ni-Fe alloy with silicon added is oxidized at 1000 °C (1800 °F), the chromia protective layer is formed on the metal, with the most external layer being a manganese- iron spinel, (MnFe)Cr 2O4. The silica is thermodynamically more stable than chromia. Moreover, the chromia dissociation pressure at the alloyCr2O3 interface is higher than the Cr-Cr2O4 equilibrium. Thus, the amount of oxygen dissolved in the alloy subsurface zone is sufficient to form an internal oxidation subscale (silica precipitation zone). Internal oxidation of chromium often affects the mechanical properties of alloys; for example, it results in a carbide-free zone. It is a result of the chromium consumption during internal oxidation and of the decomposition of less stable chromium carbides, Cr23C6 (Ref 13). In practice, the protective scales often have cracks, open porosity, or other macrodefects. Moreover, an oxidizing atmosphere can be complex, containing oxygen and sulfur compounds. The cracks are fast diffusion paths for inward transport of oxidant into the underlying steel alloy, and in such a case, the internal sulfidation by the sulfur-bearing molecules is often observed (Ref 14). Alloy Oxidation: The Doping Principle (Ref 2, 3, 4). For oxides that form according to the mechanism outlined in Ref 12 and that contain impurity cations that are soluble in the oxide, the impurities alter the defect concentration of the scale. Consequently, the oxide growth rate may also be altered by the alloy impurities. Whether the oxidation rate increases or decreases depends on the relative valences of the cations and on the type of oxide. In a p-type semiconducting oxide (NiO, CoO, MnS), the oxidation rate is controlled by cation diffusion through cation vacancies. If the number of cation vacancies can be decreased, the oxidation is slowed. If a few cations with a higher valence are substituted for the regular cations in an oxide, the vacancy concentration is increased. Adding lower-valent cations, such as lithium ions (Li+), to NiO will reduce the cation vacancy concentration. Substituting ions with the same valence as the rest of the cations in the oxide does not result in a doping effect. The n-type semiconducting oxides behave opposite to the p-type oxides. For the oxides that grow by anion diffusion through anion vacancies (typified by ZrO2), the anion vacancies exist because some cations have a lower-than-normal valence and contribute fewer electrons to the oxygen than required by the structural arrangement. Consequently, anion vacancies are present. If additional low-valent cations replace the regular higher-valent cations, the oxidation rate increases. The doping principle is not very helpful in developing oxidation-resistant alloys, because the concentration of foreign cations that can be put into solid solution in the oxide is restricted by solubility limits, and the number of foreign ions that could be used is extremely limited by their valence. Moreover, many technically important oxides, nitrides, or sulfides are ionic conductors. Selective Oxidation. An alloy is selectively oxidized if one component, usually the most reactive one, is preferentially oxidized. Otherwise, when two or more alloy elements react, the process is called concurrent oxidation. The simplest case would be a binary alloy with a uniform scale composed entirely of only the oxide that one of the components can form. An obvious example of selective oxidation would be scale formation on alloy A-B, where A is so noble that AO is not thermodynamically stable at the environmental pressure and temperature. That is, the oxygen partial pressure in the gas is less than the equilibrium (dissociation) pressure of AO oxide. Only BO scale can form. For situations in which both A and B can react with oxygen at the temperature and gas pressure involved, but A is somewhat more noble than B, the alloy composition determines which oxide forms. The scheme of selective oxidation of a binary alloy is shown in Fig. 4. The x-axis is position and the y-axis is the concentration. Alloying element A becomes depleted as it reacts at the metal-oxide interface. This creates a concentration gradient of A, causing A to diffuse from the interior of the alloy toward the surface. At the same time, depleting

A near the metal surface increases the concentration of B so that B should diffuse inward. If the diffusivities (J) of A and B in the alloy are similar to the diffusivity of A in the oxide, the element A will not be seriously depleted at the metal-oxide interface, and AO continues to form. However, the diffusion through the alloy is usually much slower than through the oxide; consequently, the concentration of B will increase at the alloy surface until it reaches , the critical concentration at which formation of BO is thermodynamically favorable. At that time, BO will form along with AO.

Fig. 4 Schematic showing the mass flow and scale growth during the selective oxidation of a binary alloy. Location is plotted on x-axis and the concentration of component A(yA) is plotted on y-axis. The initial surface of A-B alloy (at t = 0) is at x = 0. X1(t) and X2(t) are the positions of the oxide/alloy and oxidant/oxide interface, respectively. 1(t) and 2(t) are their velocities. J is the diffusivity flux. Reducing Oxidation Rates. The oxidation rate of a reactive metal cannot be markedly reduced by alloying it with a noble metal. If the concentration of the reactive element is far greater than the critical concentration needed to form an external scale, that is, if NB » , alloying with a noble metal will have very little effect on the oxidation rate. The model of parabolic, selective oxidation of a binary A-B alloy was proposed by Wagner for a binary nickelplatinum alloy (Ref 15). Wagner combined his theory of metal oxidation with the Darken model of interdiffusion in binary alloys (Ref 16), and obtained an analytical solution for the kinetics of oxidation of nickel-platinum-type alloys. His theory later became a starting point for examination of interdiffusion in p-type oxide solid solutions for non-steady-state conditions (Ref 17, 18). During selective oxidation of a binary nickelplatinum alloy, the following reaction takes place at the alloy-oxide interface: Ni(alloy) + O2(g) = NiO(s)

(Eq 11)

When diffusivity of the oxidant in the oxide is negligible and a single type of defect dominates in the metal sublattice, then the parabolic rate constant for nickel oxidation is expressed by:

(Eq 12) where and denote the equilibrium oxygen partial pressure in gas atmosphere and at the NiO-alloy interface. As a result of nickel depletion at the NiO-Ni interface, the nickel concentration in the alloy decreases. Consequently, the dissociation pressure of NiO in equilibrium with the alloy increases. Thus, is an unknown variable depending on time, and the parabolic rate is not a constant but depends on time. When the alloy is an ideal solid solution, that is, aNi = yNi, the dissociation pressure of NiO for reaction (Eq 11) can be written in the form: (Eq 13) One can relate the parabolic rate constant to the flux (J) of reacting nickel at (Ni-Pt)|NiO interface (compare Table 2 and Fig. 4). The resulting mass balance at this interface (Stefan condition) takes the form: (Eq 14) where DNiPt is the interdiffusion coefficient in the alloy at the (Ni-Pt)|NiO interface. Table 2 Basic assumptions by different authors for modeling selective oxidation of alloys Assumption

Mass transport in AXδ layer An AXδ layer is a single phase, compact, and adheres well to the alloy. Local equilibrium in the growing AXδ layer is postulated. Thus, the mass action law describes the defect structure of the growing scale. Defects in cationic sublattice dominate; the nonstoichiometry of AXδ is low. Diffusion of ionic and electronic defects is a dominating mode of mass transport. Interdiffusion in a binary alloy (Darken model) Oxidized alloy is an ideal solid solution, i.e., yi = ai An alloy molar volume does not depend on composition. Mass transport in an alloy is controlled by interdiffusion. The Darken model governs the process. Intrinsic diffusivities are equal and do not depend on composition: DA = DB = DAB = constant. Interdiffusion in multicomponent alloys Intrinsic diffusion coefficients of all elements in the system are different. Interfaces The local equilibrium at alloy-AXδ and AXδ-X2(g) interfaces prevails. Thus, it follows that aA (AXδ) = aA (alloy). The model considers stationary solution only. The nonstationary reaction period, during which the concentration of the reacting element at the alloy-AXδ

Author C. Wagner (Ref 15)

F. Gesmundo (Ref 17)

Danielewski et al. (Ref 18)

X

X

X

X

X

X

X

X

X

X

X

X

X X X

X X X

X X X

X

X

…

…

…

X

X

X

X

X …

… X

… X

interface depends on time, is allowed. On applying Eq 11, Eq 12, Eq 13, Eq 14, one can express the flux at the Ni|NiO interface as a function of the concentration of the diffusing nickel at this interface. Wagner has shown both theoretically and experimentally that alloying nickel with a small amount of platinum will have very little effect on the oxidation. Oxide growth will be markedly slower only when enough platinum is added to make nickel an extremely low concentration of the alloy. The oxidation rate is then no longer controlled by diffusion of nickel through the oxide but rather by interdiffusion of both nickel and platinum in the alloy. The practical importance of a model of selective oxidation is that it predict the ability of an alloy and/or coating to support formation of a protective scale. Composite External Scales. Wagner has shown that for an A-B alloy that is forming both AO and BO oxides, the mole fraction of B at the alloy surface must not exceed: (Eq 15) where

is the minimum concentration of A necessary to form AO, V is the molar volume of the alloy, zB is

the valence of B, MO is the atomic weight of oxygen, is the parabolic rate constant for growth of BO, and DB is the diffusion coefficient of B in the alloy. The previous formula does not take into account any complications, such as porosity and internal oxidation. If the concentration of B lies anywhere between and ), thermodynamics predicts that both AO and BO are formed. (1 − Concurrent Oxidation. When oxides of both metals form, their relative positions and distribution depend on the thermodynamic properties of the oxides and the alloy, the diffusion processes, and the reaction mechanisms. There are two common situations: both metals in a binary alloy oxidize to form two separate oxide phases, or mixed oxides, such as spinels, form. The first situation involves immiscible oxides, with the more stable oxide growing slowly. With both AO and BO stable but with rapid growth of AO and slow growth of BO, the more stable BO may nucleate first but gradually becomes overwhelmed and surrounded by fast-growing AO (Fig. 5a, b). If diffusion in the alloy is rapid, the oxidation proceeds to form an AO scale with BO islands scattered through it. However, if diffusion in the alloy is slow, the metal becomes depleted of A near the metal-oxide interface, while the growth of BO continues until it forms a complete layer, undercutting the AO (Fig. 5c). Pockets of AO at the metal-oxide interface will gradually be eliminated by the displacement reaction, AO + B(alloy) → BO + A(alloy). Because BO is thermodynamically more stable than AO, this reaction will not go in the opposite direction. Such displacement reaction continues even if the oxygen supply is cut off.

Fig. 5 Simultaneous growth of competing oxides. BO is more stable, but AO grows faster. (a) Early stage with nucleation of both oxides. (b) Later stage if diffusion in alloy is rapid. (c) Final stage if diffusion in alloy is slow The second situation involves two oxides that are partially miscible. For alloys rich in A, an AO scale will form, with some B ions dissolved substitutionally in the AO structure. If the solubility limit is exceeded when B ions continue to diffuse into the scale, BO precipitates as small islands throughout the AO layer. Even if the

solubility limit is not reached, the more stable BO may nucleate within the AO scale and precipitate. For alloys rich in B, a BO layer first forms. If B ions diffuse through the scale faster than A ions do, the concentration of A ultimately builds up in the scale close to the metal-oxide interface. An AO layer then forms underneath the BO. Double Oxides. A great deal of research has been directed toward developing alloys that form slow-growing complex oxides. The silicates are particularly important because they can form glassy structures that severely limit diffusion of ions. Therefore, silicide coatings on metals have been successfully used at high temperatures. Spinels often have extremely low diffusion rates. Spinels are double oxides of a metal with +2 valence and a metal with +3 valence, having the general formula MO · Me2O3 and also having the crystal structure of the mineral spinel (MgO · Al2O3). The iron oxide, Fe3O4, has an inverse spinel structure. For iron-chromium alloys, the spinel phase can be either stoichiometric FeO · Cr2O3 or the solid solution Fe3-xCrxO4. Although many ternary oxides tend to be brittle, much research has been devoted to minor alloy additions to improve the hightemperature mechanical properties of those ternary scales that are extremely protective. Corrosion in Complex Atmospheres. High- temperature corrosion has a wide potential for research and development, because new technologies and new materials have led to new challenges with regard to hightemperature corrosion performance. Higher operation temperatures are desired, because they result in higher efficiencies of engines and process plants. The need to operate within more aggressive environments results from the growing demand to dispose of municipal and industrial wastes and residues.

References cited in this section 1. K.P. Lillerud and P. Kofstad, J. Electrochem. Soc., Vol 127, 1980, p 2397 2. N. Birks and G.H. Meier, Introduction to High Temperature Oxidation of Metals, Edward Arnold, 1983 3. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, 1988 4. M. Schütze, in Corrosion and Environmental Degradation, Vol 1, Materials Science and Technology, A Comprehensive Treatment, R.W. Cahn, P. Haasen, and E.J. Kramer, Ed., Wiley-VCH, 1993 8. A. Nayer, The Metals Databook, McGraw- Hill, 1997 9. S. Mrowec and T. Werber, Gas Corrosion of Metals, National Center for Scientific, Technical and Economic Information, 1978 10. A. Rahmel and M. Schütze, Oxid. Met., Vol 38, 1992, p 255 11. C. Wagner, Types of Reaction in the Oxidation of Alloys, Z. Elektrochem., Vol 63, 1959, p 772–782 12. S.A. Bradford, Formation and Composition of Internal Oxides in Dilute Iron Alloys, Trans. AIME, Vol 230, 1964, p 1400–1405 13. H.W. Grünling, S. Leistikov, A. Rahmel, and F. Schubert, in Aufbau von Oxidschichten auf Hochtemperaturwerkstoffen und ihre technische Bedeutung, A. Rahmel, Ed., Deutsche Gesellschaft für Metallkunde, 1982, p 7 14. H.J. Grabke, J.F. Norton, and F.G. Castells, in High Temperature Alloys for Gas Turbines and Other Applications, W. Betz, R. Brunetaud, et al., Ed., Reidel, 1986, p 245 15. C. Wagner, J. Electrochem. Soc., Vol 99, 1952, p 369 16. L.S. Darken, Trans. AIME, Vol 174, 1948, p 184 17. F. Gesmundo and M. Pereira, Oxid. Met., Vol 47, 1997, p 507

18. M. Danielewski, R. Filipek, and A. Milewska, Solid State Phen., Vol 72, 2000, p 23


Sulfidation In distillation and gasification processes, high amounts of sulfur may be present in the form of H2S, while, at the same time, the oxygen partial pressures are extremely low. As a result, on conventional materials, sulfide scales are formed instead of oxides. Due to their high defect concentration, the sulfide scales show extremely high growth rates on these materials (Fig. 6), which is contrary to the protective effect that most of the oxide scales exhibit on these materials (Ref 19). For this reason, the use of conventional steels is basically limited to temperatures of 450 to 550 °C (840 to 1020 °F) when sulfidation occurs. To find materials solutions, intensive studies have been performed with a large number of conventional materials and with new groups of materials, such as the aluminides and the silicides. The application of coatings based on these compounds on the low-cost conventional steels seems to be feasible. The highest resistance under these conditions was by MoSi 2. However, MoSi2 failed as a coating due to spalling attributed to the difference in thermal expansion of MoSi2 and conventional steels. Iron aluminides are attractive on the basis of cost. However, again, their coefficients of thermal expansion do not match that of low-alloy, high-temperature steels, and they are sensitive to hydrogen embrittlement. Titanium aluminide (TiAl), which is applied as plasma spray coating by different techniques or in the form of a diffusion coating by codiffusion, is used for operating temperatures to 700 °C (1300 °F) in sulfidizing atmospheres.

Fig. 6 Collective plot of the temperature dependence of the sulfidation and oxidation rate of binary and ternary alloys and coatings. Source: Ref 19

Reference cited in this section 19. H. Habazaki, K. Takahiro, S.Y. Yamaguchi, K. Hashimoto, J. Dabek, S. Mrowec, and M. Danielewski, Mater. Sci. Eng. A, Vol 181/ 182, 1994, p 1099


Metal Dusting Metal dusting occurs in atmospheres with very low oxygen partial pressures and carbon activities greater than 1. The mechanisms of this process are well understood (Ref 20). During metal dusting, the material disintegrates into a powder of nanosize metal particles and graphite (graphite wool). This process is a catastrophic form of high-temperature corrosion. The incubation period for metal dusting attack in the case of high-alloy materials may last up to more than 10,000 h. The depletion of chromium in the metal subsurface zone of these materials may play a significant role in the initiation of dusting attack. In this case, the use of protective coatings is a solution. These protective coatings could contain a much higher level of the scaleforming elements. Commercial coatings are based on enriching the subsurface of steels by the diffusion of aluminum. Recent investigations show that TiAl coatings and SiAl coatings can further improve the resistance of steels under such conditions (Ref 21).

References cited in this section 20. H.J. Grabke, Mater. Corros., Vol 49, 1998, p 303 21. F. Dettenwanger, C. Rosado, and M. Schü tze, Proc. Int. Workshop on Metal Dusting, K. Natesan, Ed., Argonne National Laboratory, 2002


Chlorine Corrosion A chlorine corrosion knowledge base was created in the 1950s. The mechanisms are well understood, but the use of the metals is still limited up to 650 °C (1200 °F) for environments with low oxygen partial pressures and up to 800 °C (1470 °F) for conventional materials in oxidizing environments. The main problem when chlorine corrosion occurs is the formation of volatile metal chlorides. These metal chlorides develop partial pressures above the critical value of ~10-4 atm. Thermodynamic analysis shows that the resistant base alloy should consist of nickel with high amounts of chromium and molybdenum, because their evaporation rates would be low. However, such alloy compositions are sensitive to the presence of oxygen, because highly volatile chromium oxychlorides and molybdenum oxychlorides will be formed. Thus, the development of new materials is an important target for such environments. The existing database on the corrosion resistance under oxidizing/chloridizing environments shows that a high aluminum reservoir in the metal subsurface zone seems to be the only way to offer sufficient corrosion protection at high operating temperatures.


Summary Outline of Alloy Oxidation A schematic diagram has been constructed to show the complexity of morphologies of scale growth on binary alloys (Ref 22) (Fig. 7). The diagram illustrates some types of structures that are known to form; it does not pretend to present those that would be theoretically possible.

Fig. 7 Schematic showing the relationships between scale morphologies on binary alloys. ppn, precipitation; ppt, precipitate. Source: Ref 22 The demand for an increase in process temperatures requires new structural materials and coatings. There is still a significant potential for theoretical studies. The progress in fundamental studies may significantly speed the development of the new materials and coatings.

Reference cited in this section 22. B.D. Bastow, G.C. Wood, and D.P. Whittle, Morphologies of Uniform Adherent Scales on Binary Alloys, Oxid. Met., Vol 16, 1981, p 1–28


Acknowledgment Portions of this article have been adapted from S.A. Bradford, Fundamentals of Corrosion in Gases, Corrosion, Vol 13, ASM Handbook (formerly Metals Handbook, 9th ed.), ASM International, 1987, p 61–76.


References 1. K.P. Lillerud and P. Kofstad, J. Electrochem. Soc., Vol 127, 1980, p 2397 2. N. Birks and G.H. Meier, Introduction to High Temperature Oxidation of Metals, Edward Arnold, 1983 3. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, 1988 4. M. Schütze, in Corrosion and Environmental Degradation, Vol 1, Materials Science and Technology, A Comprehensive Treatment, R.W. Cahn, P. Haasen, and E.J. Kramer, Ed., Wiley-VCH, 1993 5. M.I. Manning, Corros. Sci., Vol 21, 1981, p 301 6. W. Christl, A. Rahmel, and M. Schütze, Oxid. Met., Vol 31, 1989, p 1 7. P. Hancock and R.C. Hurst, The Mechanical Properties and Breakdown of Surface Films at Elevated Temperatures, Advances in Corrosion Science and Technology, Vol 4, R.W. Staehle and M.G. Fontona, Ed., Plenum Press, 1974, p 1–84 8. A. Nayer, The Metals Databook, McGraw- Hill, 1997 9. S. Mrowec and T. Werber, Gas Corrosion of Metals, National Center for Scientific, Technical and Economic Information, 1978

10. A. Rahmel and M. Schütze, Oxid. Met., Vol 38, 1992, p 255 11. C. Wagner, Types of Reaction in the Oxidation of Alloys, Z. Elektrochem., Vol 63, 1959, p 772–782 12. S.A. Bradford, Formation and Composition of Internal Oxides in Dilute Iron Alloys, Trans. AIME, Vol 230, 1964, p 1400–1405 13. H.W. Grünling, S. Leistikov, A. Rahmel, and F. Schubert, in Aufbau von Oxidschichten auf Hochtemperaturwerkstoffen und ihre technische Bedeutung, A. Rahmel, Ed., Deutsche Gesellschaft für Metallkunde, 1982, p 7 14. H.J. Grabke, J.F. Norton, and F.G. Castells, in High Temperature Alloys for Gas Turbines and Other Applications, W. Betz, R. Brunetaud, et al., Ed., Reidel, 1986, p 245 15. C. Wagner, J. Electrochem. Soc., Vol 99, 1952, p 369 16. L.S. Darken, Trans. AIME, Vol 174, 1948, p 184 17. F. Gesmundo and M. Pereira, Oxid. Met., Vol 47, 1997, p 507 18. M. Danielewski, R. Filipek, and A. Milewska, Solid State Phen., Vol 72, 2000, p 23 19. H. Habazaki, K. Takahiro, S.Y. Yamaguchi, K. Hashimoto, J. Dabek, S. Mrowec, and M. Danielewski, Mater. Sci. Eng. A, Vol 181/ 182, 1994, p 1099 20. H.J. Grabke, Mater. Corros., Vol 49, 1998, p 303 21. F. Dettenwanger, C. Rosado, and M. Schü tze, Proc. Int. Workshop on Metal Dusting, K. Natesan, Ed., Argonne National Laboratory, 2002 22. B.D. Bastow, G.C. Wood, and D.P. Whittle, Morphologies of Uniform Adherent Scales on Binary Alloys, Oxid. Met., Vol 16, 1981, p 1–28

A. Gil, Methods for Measuring Gaseous Corrosion Rates, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 115–116

Methods for Measuring Gaseous Corrosion Rates Aleksander Gil, University of Mining and Metallurgy, Cracow, Poland

Introduction CORROSION RATES of metal alloy test specimens having identical masses but different shapes (e.g., flat circular, cylindrical, or spherical), or identical shapes but different sizes, are not necessarily the same. It has been shown that the best reproducibility of results is obtained with flat samples (Ref 1). Therefore, corrosion studies usually use rectangular- or circular- shaped flat specimens, with thicknesses not exceeding 1 mm (0.04 in.) and total surface area of a few square centimeters (approximately 1 in.2). To eliminate the effect of edges and corners on the corrosion process, these sample parts can be rounded. The corrosion process is additionally

influenced by surface finish (grinding, mechanical or electrolytic polishing, etc.). It is advised to use 600-grit emery paper and ultrasound cleaning when preparing test specimens (Ref 2).

References cited in this section 1. J. Romański, Corros. Sci., Vol 8, 1968, p 67s–89s 2. H.J. Grabke, W. Auer, M.J. Bennett, F. Bregani, F. Gesmundo, D.J. Hall, D.B. Meadowcroft, S. Mrowec, J.F. Norton, W.J. Quadakkers, S.R.J. Saunders, and Z. Zurek, Werkst. Korros., Vol 44, 1993, p 345

A. Gil, Methods for Measuring Gaseous Corrosion Rates, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 115–116 Methods for Measuring Gaseous Corrosion Rates Aleksander Gil, University of Mining and Metallurgy, Cracow, Poland

Discontinuous Methods Most metals react with oxidizing gases (oxygen, nitrogen, sulfur, chlorine) to yield solid products such as scale. These scales are generally adherent to the metallic substrate in a wide temperature range. In such cases, any direct and continuous records of weight losses are impossible. The corrosion rate data can be obtained indirectly by measuring: • • • • •

Scale thickness Scale weight per unit surface area Loss of metal thickness Loss of material weight per unit surface area Weight change of oxidant bonded in the scale per unit surface area as a function of time

Thickness or weight of the scale or the uncorroded material can be measured in a discontinuous manner at room temperature. Specimens with identical starting dimensions can be exposed to a controlled environment and then removed and measured at differing exposure times. This series of experiments and samples provides the time dependence of the corrosion process. Scale thickness is measured on the metallographic cross sections. The procedure is simple when the scale is homogeneous, has the same composition on the whole surface, and when internal oxidation does not occur in the metallic core. The dependence of scale thickness on time can be used for the determination of a kinetic law for the oxidation rate. In the case when the rate-controlling step in the oxidation process is the diffusion of ions through a compact barrier layer of oxide with the chemical potential gradient as the driving force, the parabolic rate law is usually observed. As the oxide grows thicker, the diffusion distance increases, and the oxidation rate slows down. The rate (dx/dt) is inversely proportional to the oxide thickness (x) or:

On integration, the parabolic equation is obtained: x2 = 2kpt and the parabolic rate constant, kp, can be determined from the experiment in units of cm2/s or similar length squared per time units.

To measure the weight of the scale or the uncorroded metal, it is necessary to remove the corrosion product, which is a rather troublesome task. There are standard practices for removing these corrosion products, however (Ref 3). The discontinuous methods for examining uncorroded base metals are rarely used in studying high-temperature corrosion. However, it should be stressed that the discontinuous methods have certain advantages. Discontinuous methods allow for examination of specimens after different lapse times, so it is possible to follow changes in scale and alloy structure at different stages of the corrosion process.

Reference cited in this section 3. “Preparing, Cleaning, and Evaluating Corrosion Test Specimens,” G 1, ASTM International


Continuous Methods Direct measurements of weight or thickness loss during the corrosion process are possible only in the case when the corrosion product is volatile. Metals that form volatile oxides are not numerous. For example, tungsten, molybdenum, and platinum belong to this group. In such cases, the rate of corrosion can be determined continuously on the basis of thickness or weight changes in time. Thickness can be measured during the corrosion experiment by means of a cathetometer. The weight of an oxidant bonded by a unit surface area, Δm/A, in units of g/cm2, can be measured continuously, and in such a case, oxidation of a single sample provides information for determining the rate of corrosion. Again, if the kinetic law of this process is parabolic, the weight-based parabolic rate constant, kp, can be calculated in units of g2/cm4 · s. The three principal methods are volumetric, manometric, and thermogravimetric. The volumetric method is based on measuring the volume of gaseous oxidant consumed by the corroding material under isobaric conditions (Ref 4). This method is highly sensitive but has some limitations. Oxidation cannot be performed in gas mixtures, and gaseous products cannot be formed, because they would disable accurate measurements of oxidant volume. The manometric method consists of measuring the pressure variations in the reaction zone brought about by the oxidation process (Ref 5). The manometric method is as sensitive as the volumetric one and has similar limitations. An additional problem is that it does not allow maintaining strictly isobaric conditions during the oxidation test. Thermogravimetric Analysis (TGA). This method is most frequently used in oxidation rate measurements. It is experimentally simple and does not impose any limitations as far as the gas phase composition is concerned. Thermogravimetry consists of measuring the weight changes of the oxidized sample placed in the reaction zone (Fig. 1a). The precision of new thermobalances is such that mass changes as low as 0.1 μg can be measured. This enables determining very small parabolic rate constants, on the order of 10-14 g2/cm4 · s. Gaseous reaction products do not affect the accuracy of measurements, although the rate of reaction calculated from experimental curves may differ from the “real” rate.

Fig. 1 Schematic drawing of thermobalances for thermogravimetric analysis. (a) Tube furnace and microbalance without compensation for hydrostatic lift. (b) With compensation for hydrostatic lift The gaseous components may appear as a result of direct reaction between metal and oxidant, but they may also appear as products of secondary reaction between the primary product and the oxidant. In both cases, the weight change taken from the TGA measurement, Δm(t), is a difference between the weight of the oxidant taking part in the reaction with the metal, Δmox(t), and the weight of a volatile component of the corrosion product, Δmvol(t): Δm(t) = Δmox(t) - Δmvol(t) The real corrosion rate in such a case is higher than that resulting from the analysis of the TGA curve. This situation arises in chromia formers when the oxidation temperature exceeds approximately 900 °C (1650 °F). Chromium (III) oxide, Cr2O3, oxidizes to a volatile CrO3, and in the presence of water vapor, other volatile compounds are formed, such as CrO2(OH)2 (Ref 6). An important source of measuring errors in the TGA method, especially when the weight gains are small, may come from the weight changes of the fixture that supports the sample. The suspensions are usually made of thin

platinum or platinum-rhodium wires, quartz fibers, or preoxidized Fe-Cr-Al wires covered with a protective film of alumina. High temperatures (over 1000 °C, or 1830 °F) and at high partial pressures of oxygen combined with prolonged test times (thousands of hours) would cause a volatile PtO 2 to form. Platinum wires should be avoided. At low partial pressures of oxygen and high temperatures, using quartz suspensions is not recommended because of possible formation of a volatile oxide, SiO. However, by weighing the suspension fixture before and after the experiment, it can be checked whether or not the weight change of the fixture due to corrosion is negligible. In TGA measurements of oxidation rate, it should be realized that the measured weight of the sample also includes that of gases that are dissolved in the metallic phase without taking part in the oxidation process. One metal with high solubility for oxygen is titanium. Its high- temperature variety, α-Ti, can dissolve as much as 30 at.% oxygen. In the TiAl alloys under the protective alumina scale, there appears a Z- phase, Ti5Al3O2 (Ref 7), which may bond similar amounts of oxygen as the protective scale. The interpretation of thermogravimetric curves is easier when scale morphologies and cross sections are also examined. This is particularly important when the scale has varying composition on the metal surface and when the internal oxidation zone appears in the metal substrate. When the specimen is inserted from room temperature into the hot zone of the furnace, it should be realized that its weight slightly increases as a result of hydrostatic lift. The density of gas is inversely proportional to temperature. The hydrostatic lift should be taken into consideration in a situation when the total weight gain of the sample is expected to be small. The effect can be eliminated by means of a thermobalance where the second arm is loaded with a sample having the same dimensions as the investigated one but made of an inert material, for example, Al2O3 (Fig. 1b). Both samples are heated to the same temperature. Thermogravimetric curves obtained for the same material in different laboratories exhibit a significant scatter caused by differences in measuring instrument design and experimental conditions. The discrepancies may be due to different gas flow rates in the reaction zone of the furnace, variations in the atmospheric pressure, different air humidity in the case of open systems, or slight differences in temperature or temperature fluctuations in the furnace. Because of simplicity and high accuracy of measurements, the TGA method is the one most often encountered in corrosion studies. However, other methods can also be used successfully to suit some cases. The thickness of very thin, transparent oxide films is determined by measuring the light interference or polarization changes. It is also possible to determine the rate of corrosion on the basis of measured electrical conductivity of the corrosion product.

References cited in this section 4. J. Engell, K. Hauffe, and B. Ilschner, Z. Elektrochem., Vol 58, 1954, p 478 5. W. Cambell and W. Thomas, J. Electrochem. Soc., Vol 91, 1939, p 623 6. C.A. Stearns, F.K. Kohl, and G.C. Fryburg, J. Electrochem. Soc., Vol 121, 1974, p 89 7. N. Zheng, W. Fischer, H. Gruebmeir, V. Shemet, and W.J. Quadakkers, Scr. Metall. Mater., Vol 33, 1995, p 47


References

1. J. Romański, Corros. Sci., Vol 8, 1968, p 67s–89s 2. H.J. Grabke, W. Auer, M.J. Bennett, F. Bregani, F. Gesmundo, D.J. Hall, D.B. Meadowcroft, S. Mrowec, J.F. Norton, W.J. Quadakkers, S.R.J. Saunders, and Z. Zurek, Werkst. Korros., Vol 44, 1993, p 345 3. “Preparing, Cleaning, and Evaluating Corrosion Test Specimens,” G 1, ASTM International 4. J. Engell, K. Hauffe, and B. Ilschner, Z. Elektrochem., Vol 58, 1954, p 478 5. W. Cambell and W. Thomas, J. Electrochem. Soc., Vol 91, 1939, p 623 6. C.A. Stearns, F.K. Kohl, and G.C. Fryburg, J. Electrochem. Soc., Vol 121, 1974, p 89 7. N. Zheng, W. Fischer, H. Gruebmeir, V. Shemet, and W.J. Quadakkers, Scr. Metall. Mater., Vol 33, 1995, p 47

R.A. Rapp, Corrosion by Molten Salts, Corrosion: Fundamentals, Testing, and Protection,, Vol 13A, ASM Handbook, ASM International, 2003, p 117–123

Corrosion by Molten Salts Robert A. Rapp, The Ohio State University

Introduction THE OPERATION of high-temperature engineering systems, despite their associated materials problems, is inherent to advanced technologies that strive to gain an advantage in thermodynamic driving force or in reaction kinetics or both. Certain high-temperature systems involve the contact of materials with a large quantity, or significant depth, of a salt solution above its liquidus temperature. Such systems include the molten carbonate fuel cell (MCFC), molten chloride baths to melt used aluminum beverage cans, the Hall-Heroult aluminum electrowinning cell with a fluoride (cryolite)-base electrolyte, nitrate/nitrite heat-exchange salts, and fused salt descaling or heat treatment baths. In other important engineering systems, the accelerated corrosion of materials results from the contact of materials with thin films of deposited fused salts. This important corrosion mode, analogous in certain aspects to aqueous atmospheric corrosion near room temperature, is called hot corrosion. Gas turbines, steam generators, incinerators, and various petrochemical process vessels operate at high temperatures (600 to 1100 °C, or 1100 to 2000 °F) and involve metallic or ceramic materials in contact with combustion product gases containing inorganic impurities. As these gases are cooled below their dewpoint, fused alkali sulfate-base salt films may condense on the hardware to generate an aggressive hot corrosion condition. Sometimes, the salt is deposited directly as liquid droplets from the gas stream, sometimes as solid salt particles shed from an upstream filter or compressor for a marine gas turbine. Both in understanding and testing for corrosion by molten salts at high temperatures, the relevant salt depth and the gaseous atmosphere must be respected. The chemistry and corrosion have been studied for many fused salt systems: chlorides, fluorides, carbonates, sulfates, hydroxides, oxides, and nitrate/nitrite. In analog to aqueous solutions, each of these melts is a dominant ionic conductor of electricity (an electrolyte), and each melt exhibits both an oxidation potential and an acid/base character, which depend on the salt composition and its surrounding environment. Because of their high thermodynamic stability, fused alkali sulfates are frequently formed from the combustion of fossil fuels,

because impurities from both the fuel and the combustion air (perhaps containing a sea salt aerosol) react to form a product based on fused sodium sulfate. Besides its specific engineering significance, fundamental studies of the sodium sulfate system are available in the greatest depth, so this system is chosen to illustrate the important aspects of fused salt corrosion in this article. However, comments pointing out particular aspects of other salt systems are also made. Reference 1 provides a description of further engineering aspects of fused salt corrosion.

Reference cited in this section 1. G.Y. Lai, High-Temperature Corrosion of Engineering Alloys, ASM International, 1990

R.A. Rapp, Corrosion by Molten Salts, Corrosion: Fundamentals, Testing, and Protection,, Vol 13A, ASM Handbook, ASM International, 2003, p 117–123 Corrosion by Molten Salts Robert A. Rapp, The Ohio State University

Phase Stability and Salt Chemistry In many respects, the corrosion of materials by fused salts at high temperature has many common aspects with aqueous corrosion near room temperature. Because of its high thermodynamic stability, fused sodium sulfate is frequently the aggressive salt film in combustion product gases that causes hot corrosion. At 1200 K (1700 °F) for the equilibrium: Na2SO4 = Na2O + SO3(g)

(Eq 1) (Eq 2)

where is the thermodynamic activity for Na2O, and P is the partial pressure of SO3. According to Ref 2, one can define the basicity of a sodium-sulfate-base melt as log and its acidity as log . In analog to the Pourbaix approach to the study of aqueous corrosion, Fig. 1 presents a phase stability diagram for the Na-S-O system at 1173 K (1652 °F), where the ordinate is the oxidizing potential and the abscissa is a quantitative measure of the melt basicity/acidity (Ref 3). Figure 1 is calculated from known values of the standard Gibbs formation energies for the phases indicated, and the dotted lines partition the field of Na2SO4 stability according to the dominant minority solute species in the Na2SO4 solvent. In contrast to aqueous solutions, molecular oxygen has a very low solubility in fused Na2SO4, and dissolved SO3 is the primary oxidizing solute (Ref 4). Oxygen molecules are also almost insoluble in other fused salts. Therefore, in contrast to aqueous solutions, the hydrogen ion has little presence or importance in most fused salts; instead, the electrochemical reduction of oxyanion species is usually important, and the alkali oxide activity is the key indication of melt basicity.

Fig. 1 Na-S-O phase stability diagram for 1173 K (1652 °F). Source: Ref 3 Following the early authors of fused salt stability diagrams, the abscissa for Fig. 1 increases to the right with increasing melt acidity, the opposite to the Pourbaix formalism for aqueous solutions. The basicity and acidity, respectively, of other fused salt systems can be described analogously, so long as only one cation is involved: Na2CO3 (log , log ); NaOH (log , log ); and NaNO3 (log , log ). Similar Pourbaix-style phase stability diagrams can be calculated for the other salt systems; today, readily available software will provide such plots. Electrochemical probes have been developed for the individual measurement of the thermodynamic activities of oxygen and sodium in fused sodium-sulfate-base melts (Fig. 2). These probes involve Na+- and O2--conducting solid electrolytes in combination with reference electrodes of known, fixed activities for sodium and oxygen (Ref 3, 5, 6). By the combined use of these two sensors, the thermodynamic activity of Na 2O can be determined, providing a quantitative measure of the melt basicity (abscissa scale in Fig. 1) equivalent to the use of the pH electrode for aqueous solutions. The oxygen sensor alone provides the measurement of the oxygen potential, corresponding to the redox ordinate scale of Fig. 1. When only sodium cations are involved, with sulfate or several other anions, these probes can accurately provide a measure of melt basicity defined as log and the oxidation potential (log ). Analogous sensors have been used with fused carbonate melts, fused chloride melts, and for oxyfluoride melts (Ref 7), and they could be developed for other fused salt systems.

Fig. 2 Experimental electrochemical reference electrodes to simultaneously measure sodium and oxygen activities and thereby, melt basicity. (a) Ag/Ag+ electrode. (b) ZrO2 electrode. Source: Ref 3, 5, 6 Because of the high temperatures involved, the usual highly oxidized state of the aggressive fused salts, and the thermodynamic incompatibility of the oxidizing salt and the reducing metal, the hope for corrosion immunity

for materials cannot be realized. Rather, to achieve passivity, the stability and solubility of a protective oxide in contact with the corrosive medium are of great importance, as is the consequence of a localized failure/penetration of the protective oxide. In this context, a protective oxide refers to one obeying diffusionlimited growth kinetics. Thus, the engineer must choose materials and modify processes or environments to achieve an adherent protective oxide with minimal solubility in the fused salt. Following the Pourbaix approach, the solubility of a (protective) oxide in a (corrosive) solvent is understood by the superposition of a phase stability diagram for that oxide over that for the solvent, for example, Fig. 1 for the Na2SO4 salt. Thus, quaternary phase stability diagrams have been generated for many M-Na-S-O systems for the metals of interest, as illustrated by the Cr- Na-S-O stability diagram of Fig. 3 (Ref 8). As for aqueous Pourbaix diagrams, within the field of Cr2O3 stability, dashed lines indicate the calculated concentrations for each solute (assuming an ideal solution). The ratio between the actual measured solubilities and the calculated solubilities provides a value for the activity coefficient for the particular solute. Implicit to Fig. 3 is the absence of any field for the stability of chromium metal, either pure or in an alloy. Thus, anytime the fused Na 2SO4 salt would contact chromium or an alloy containing chromium, a reaction must be expected, most usually to form a sulfide product phase. This behavior is also common to the other important base-metal components, iron, nickel, cobalt, aluminum, and so on (Ref 6). The most important aspect of alloy protection is to prevent contact of the alloy and the salt, for example, by an adherent protective oxide with minimal solubility.

Fig. 3 Na-Cr-S-O phase stability diagram for 1200 K (1700 °F). Source: Ref 8 The Cr-Na-S-O system of Fig. 3 is rather complicated, because Cr2O3 can form two acidic (Cr2(SO4)3 and CrS) and two basic (Na2CrO4 and NaCrO2) solutes. The experimentally measured Cr2O3 solubility at 1200 K (1700 °F) and 1 atm O2 and the relevant dissolution reactions with the predicted slopes are presented in Fig. 4 (Ref 8). In this range of dilute solutions, the expected dependencies for each of the three solutes are closely followed, consistent with Fig. 3. Additional experiments at lower showed that NaCrO2 becomes the dominant basic solute instead of Na2CrO4. Whenever a given cation from an oxide changes its valence on dissolution, for example, forming from Cr2O3 dissolution, the resulting solubility depends on both the basicity and the oxygen activity. In this instance, the solubility of Cr2O3 is higher at high oxygen activity and lower at low oxygen activity, for a given basicity. The experimental results of Fig. 4 are in complete agreement with those expected by the Pourbaix thermodynamic approach of Fig. 3, except that at such a high temperature, the oxide solubility can actually be measured and applied to a corrosion model.

Fig. 4 Experimentally measured solubilities for chromium oxide in fused Na2SO4 at 1200 K (1700 °F) and 1 atm oxygen. Source: Ref 8 Figure 5 presents a compilation of such experimentally established solubility plots for 1200 K (1700 °F) and 1 atm O2 (Ref 5, 6). From these plots, the oxides for nickel and cobalt are the most basic, and the oxides of chromium and aluminum are the most acidic, with approximately six decades of basicity separating their respective solubility minima. For every oxide except SiO2, the solubility curves comprise the individual contributions from simple (uncomplexed) acidic and basic solutes, with readily predicted slopes. Each plot (except SiO2) then is comprised of a leg plotting the basic solubility (left side) and another leg plotting the acidic solubility (right side). For basic solubility: (Eq 3) for acidic solubility: (Eq 4)

Fig. 5 Compilation of measured solubilities for several oxides in fused pure Na2SO4 at 1200 K (1700 °F). Source: Ref 5, 6 Silica, in the indicated (acidic) experimental range, produces only a molecular (no ionic) solute, so no basicity dependence was found. (However, a basic solute does form for silica in more basic solutions.) The solubility plots of Fig. 5 offer some insight concerning the selection of materials for specific applications. As the first cut, one should choose a material whose slow-growing, adherent oxide scale is close to its solubility minimum for the local conditions in an application, as determined by calculation or measurement. Accordingly, a silica protective film is known to be very resistant to acidic salts, for example, on SiC or Si3N4 materials in high-sulfur coal environments (Ref 9). In contact with a quite basic salt, for example, alkali carbonate, silica is readily attacked to form a soluble Na2SiO3 product. So, perhaps counter to intuition, combustion product gases in a high-sulfur (acidic) coal-burning facility are less corrosive to a silica-protected material than in an environment with low sulfur but high alkali content. In gas turbine practice, coatings based on Cr2O3 and Al2O3 have proven effective, because the very acid service condition resulting from sulfur in the fuel and sulfate from seawater corresponds to the minima for their solubility plots. While these protective oxides are the best for this application, any penetration of the oxide provides the condition needed to form sulfides for all of the superalloy components. For this reason, the trace reactive-element additions, such as yttrium, cerium, and lanthanum, which are added to high-temperature alloys to improve scale adherence, also contribute to protection from hot corrosion.

References cited in this section 2. D. Inman and D.M. Wrench, Br. Corros. J., Vol 1, 1966, p 246 3. C.O. Park and R.A. Rapp, J. Electrochem. Soc., Vol 133, 1986, p 1636 4. R. Andresen, J. Electrochem. Soc., Vol 126, 1979, p 328

5. R.A. Rapp, Corrosion, Vol 42, 1986, p 568 6. Y.S. Zhang and R.A. Rapp, J. Met., Vol 46, 1994, p 47 7. Y.S. Zhang, X. Wu, and R.A. Rapp, Metall. Mater. Trans. B, Vol 34, 2003, p 235 8. Y.S. Zhang, J. Electrochem. Soc., Vol 133, 1986, p 655 9. N.S. Jacobson, J. Am. Ceram. Soc., Vol 76, 1993, p 33


Electrochemistry and Fused Salt Corrosion The electrochemical tests that have proven so important to understanding aqueous corrosion, as reviewed elsewhere in this Volume, also apply directly to corrosion by fused salts. However, fused salt experimentation is probably made more difficult and complicated by the myriad salts of interest. Again, it is important to differentiate between a deep melt and a surface film, because certain important differences in mechanistic interpretation apply. Likewise, because the acid/base and oxidizing potential of a fused salt may be decided by an equilibrium with an ambient gas phase, it is important to match the experimental environment with the service conditions, to the extent possible. For this reason, sulfur-free air should not be substituted in the laboratory testing of hot corrosion for sulfur- laden combustion product gases. Relative to the laboratory electrochemical methods, all the traditional and more advanced techniques can be used for deep melts (crucible tests), for example, three-electrode polarization scanning to establish the kinetics for cathodic and anodic halfcell reactions, chronopotentiometric and chronoamperometric scans, cyclic voltammetry, impedance spectroscopy, electrochemical noise analysis, and so on. Likewise, potentiometric monitoring of the corrosion potential can be done, even for thin fused salt films. Details for such methodologies are beyond the scope of this article, but example studies that establish the electrochemical reactions are available for pure Na2SO4 (Ref 3, 10, 11). The Cr-Na-S-O system (Ref 12) and the Mo-Na-S-O system (Ref 13) have been studied in detail. The mechanistic information that can be gleaned for these electrochemical studies, at least for fused sulfate corrosion of metals, points out the special importance of the cathodic reduction reaction. As shown in Fig. 6, on deep cathodic polarization for a platinum working electrode in fused Na2SO4, the reduction of the dissolved (in an acidic melt) becomes quickly limited by diffusion of the ion to the cathode. A resulting oxidant sharp drop in the oxidizing potential of the platinum then causes the reduction of the sulfate ion (Ref 3): (Eq 5) although this reaction probably occurs over several intermediate steps. Potentiometric reference electrodes, as shown in Fig. 2, have established that when the oxidizing potential dropped rapidly, sulfide ions were formed, and the melt in contact with this cathode greatly shifted in a basic direction.

Fig. 6 Trace of basicity and oxygen activity measured on polarization of a platinum working electrode in 0.1% SO2-O2 gas at 900 °C (1650 °F). Scan rate, 0.1 mV/s. Source: Ref 3 The relevance of this laboratory experiment on platinum to the corrosive attack of a base metal is apparent. A bare engineering alloy (Fe- Ni-Cr-Al) at high temperatures constitutes a highly reducing condition, equivalent to a deep cathodic polarization driven electrically at the platinum working electrode. If the reactive salt should come into contact with the bare alloy, then immediate sulfidation must be expected, and the salt would locally become very basic. With eutectic temperature in the nickel-sulfur system at 645 °C (1195 °F), perhaps a liquid sulfide product would be formed. This is known to wet and progress along grain boundaries of the alloy. In contact with fused Na2SO4, because both chromium and aluminum form thermodynamically more stable sulfides than nickel or iron, these more reactive alloying elements could be removed from the alloy by precipitation of internal sulfides. However, because chromium and aluminum are usually intended to grow and maintain the protective oxide scale, their reaction to form sulfides would eliminate this protection. Because fused salts are known to readily wet surfaces and interfaces, local ingress of the salt through flaws in the protective scale may also lead to detachment of the scale. Finally, the site of such local sulfide formation would become much more basic than the bulk melt (or salt film), and thus, a gradient in the solubility for the protective oxide in the salt should be established. The significance of the local basicity shift is detailed later. The extreme consequences of a sulfate melt contacting the bare alloy suggests some important aspects for experimental corrosion testing. First, one can never simply immerse a bare base-metal specimen into a bath of hot sulfate salt. An immediate reaction must occur (the metal and the salt are thermodynamically incompatible), so no eventual protection can be expected. Both in experimentation and in engineering applications, some intentionally formed protective oxide should be provided by some preoxidation (in the absence of sulfur). The resistance to fused salt corrosion would then be measured by the initiation time required for the salt to selectively attack some chemical or physical flaw in the protective scale, which would lead to the salt contacting the alloy substrate. For many salt and alloy combinations, rapid attack leading to a failure must be expected after the initiation time. Of course, different salts pose different types and levels of attack. Consider a nickel-base containment vessel for a fused nitrate/nitrite or carbonate salt. Because nickel does not form a stable nitride or carbide corrosion product, a severe reaction equivalent to sulfide formation in sulfate corrosion is not experienced. However, metal-salt contact would still provide an electrochemical reduction of the oxidizing salt, with a corresponding increase in local melt basicity. If instead, a nickel-chromium or iron-chromium alloy were involved, a nitride or carbide of chromium could result on salt-metal contact. When chloride salts, or complex salts containing chloride, contact a metal, having breached the protective oxide, volatile species are often formed and lost from the alloy. The alloy components iron and chromium often suffer evaporation attack via loss of volatile chlorides, which requires a porous scale in order for the vapors to escape. Sometimes, after volatile species have left the alloy reaction site, the metal component of the vapor species may be reoxidized away from this interface, reforming chlorine (or HCl) vapor to return the substrate to repeat the reaction and continue the cycle.

Such a problem is experienced in incinerator gases and chloride-containing fly ash deposits from coal combustion (Ref 14).

References cited in this section 3. C.O. Park and R.A. Rapp, J. Electrochem. Soc., Vol 133, 1986, p 1636 10. W.C. Fang and R.A. Rapp, J. Electrochem. Soc., Vol 130, 1983, p 2335 11. Y.M. Wu and R.A. Rapp, J. Electrochem. Soc., Vol 138, 1991, p 2683 12. D.Z. Shi and R.A. Rapp, Werkst. Korros., Vol 41, 1990, p 215 13. J. Kupper and R.A. Rapp, Werkst. Korros., Vol 38, 1987, p 187 14. M. Spiegel, Mater. High Temp., Vol 14, 1998, p 221


Scale Fluxing: A Hot Corrosion Mechanism Such a detailed knowledge of oxide solubilities in fused Na2SO4, as shown in Fig. 5, leads to the idea that sustained fused salt corrosion could result upon dissolution of the protective oxide and reprecipitation of that solute elsewhere in the salt at a site of lower solubility. The dissolution attack of a protective oxide would reduce its thickness and ultimately lead to its localized attack and scale penetration. A reprecipitation of the solute to form separate oxide particles in the salt would not contribute to any protection. This scale fluxing mechanism is consistent with observed corroded microstructures (porous oxide in a salt matrix) and is applicable to either a deep salt melt or a thin salt film hot corrosion. This scale fluxing mechanism is illustrated for a thin salt film in Fig. 7 (Ref 15) and is denoted as the negative solubility gradient criterion. The decrease in solubility for the oxide across the salt film could be maintained by a combination of gradients in the local salt basicity and its oxygen activity. Probably, such gradients would be more extreme for a salt film (hot corrosion) than for a deep melt. (Remember that the oxidant that is soluble in fused sodium sulfate is dissolved SO 3 [

ions] and not molecular O2, as for aqueous solutions.)

Fig. 7 Reprecipitation of porous MO oxide supported by a negative solubility gradient in the fused salt film. Source: Ref 15 Figure 8 illustrates several conditions leading to a negative solubility gradient on the basis of an idealized solubility curve for the protective oxide (Ref 15). In case A, sustained basic fluxing and reprecipitation is favored, because a dominant basic solute has a higher solubility at the oxide-salt interface than at the salt-gas interface. In case C, an acidic solute satisfies these conditions. In case B, a negative solubility gradient is realized whenever the local conditions cause the corresponding interfacial solubilities to straddle a minimum. Other factors leading to concentration (and solubility) gradients in the salt film include preferential evaporation of some component from the salt-gas interface, such as oxide or oxychloride vapors of refractory metals: vanadium, molybdenum, tungsten, and so on. Reactions, for example, sulfidation, between the salt and the metal lead to high basicity at the metal- salt interface. Likewise, the basicity at the site of the electrochemical reduction reaction is always increased. However, as detailed in Ref 5 and 15, the cathodic reduction reaction might occur at either side of the salt film, depending on several factors, for example, the presence of multivalent transition metal ions in solution. A temperature gradient in the salt at the interface would also contribute to a solubility gradient. Certainly, the solubility of any oxide in any salt would decrease with decreasing temperature. The several sorts of reactions and a thermal gradient all generally lead to some gradient in the solubility of the protective oxide scale. The magnitude of the problem would be greater for a thin salt film than for a deep melt. Accordingly, laboratory studies to test or clarify hot corrosion should not be conducted in a deep melt. The most representative, but expensive, testing for hot corrosion is provided by burner rigs supporting salt injection, where the fuel has the representative sulfur content.

Fig. 8 Types of sustained hot corrosion of a pure metal. Site I is the oxide-salt interface, and site II is the salt-gas interface. Source: Ref 15 Historically, it has been demonstrated (Ref 16) that a condensed fused salt film is required for severe hot corrosion attack, and that Na2SO4 vapor in air is innocuous. Researchers (Ref 17, 18, 19) showed that the attack does not necessarily depend on an alloy sulfidation-oxidation sequence but rather on a destructive dissolution (fluxing) of the normally protective oxide in dependence on the presence of Na2O in the fused salt film. Thin coatings of Na2CO3 and NaNO3 on the alloys produced accelerated attack similar to Na2SO4, although sulfur was absent from the system. Finally, while pure nickel reacts rapidly with Na2SO4, because a sulfide phase is formed, a Na2CO3 film does not produce such rapid attack of nickel, because no carbide is formed. In total, these studies illustrated qualitatively that salt-alloy reactions can (but might not) shift the acid/base character of a fused salt film, thereby effecting a fluxing of the protective scale. During nominally the same time period, researchers (Ref 20) studied the attack of pure nickel and nickel-base alloys beneath a fused Na2SO4 film at 1000 °C (1830 °F) and suggested a scale fluxing mechanism consistent with Fig. 7. A basic fluxing of the protective NiO scale and subsequent oxide reprecipitation resulted from the following reaction: 4Ni + Na2SO4 = Na2O + 3NiO + NiS

(Eq 6)

According to this reaction, nickel metal reacts with fused Na2SO4 (not a gas phase reactant) to form NiO and thereby reduces the oxygen activity until the sulfur activity of the salt contacting the metal is sufficient to form liquid nickel sulfide. As sulfur is removed from the salt by reaction with nickel, the local salt basicity (sodium oxide activity) increases, such that the normally protective NiO scale is dissolved/fluxed to form a basic solute, proposed to be nickelate ions. As part of the characteristic morphology for a corroded nickel specimen, nickel sulfide was seen in the metal, and nonprotective NiO particles were found precipitated within the external salt film. These results supported what is today known as a basic fluxing reaction, that is, corrosive attack by forming a basic solute of the protective scale. Researchers (Ref 21) showed that Al2O3- forming nickel-base alloys containing the refractory elements molybdenum, tungsten, or vanadium can suffer Na2SO4-induced catastrophic attack, with the formation of , , or ions, respectively). Again, the stable solute ions involving oxide ions ( precipitation of a nonprotective oxide (Al2O3) in the salt film correlated to a local reduction in the Al2O3 solubility as MoO3 and WO3evaporation occurred at the salt-gas interface. Later, researchers (Ref 22) showed that the acidic solubility of any oxide (except SiO2) would be increased by the presence of strongly acidic components in the salt. This type of accelerated attack associated with refractory metal oxide solutes has become known as acidic fluxing. Researchers (Ref 23) also demonstrated acidic fluxing in the attack of superalloys containing refractory-metal elements.

Besides the designations of basic and acidic scale fluxing, the subject of hot corrosion is otherwise classified into type I high-temperature hot corrosion (HTHC) and type II low-temperature hot corrosion (LTHC). For type I HTHC, the salt film is considered to be the relatively pure solvent, and the temperature must nominally exceed its melting point. For type II LTHC, the liquidus temperature for the salt film could be significantly lower than the melting point for the pure solvent because of a significant dissolution of some corrosion product. The dissolution reactions previously discussed here generally refer to type I HTHC. Researchers (Ref 24) showed that the corrosion kinetics of Na2SO4-coated coupons of Ni-30Cr and Co-30Cr alloys suffered maximum rates at approximately 650 and 750 °C (1200 and 1380 °F), respectively, that is, well below the 884 °C (1623 °F) melting point for the pure Na2SO4 The reaction product morphologies were characterized by a nonuniform attack in the form of pits, with only minor sulfide formation and little depletion of chromium or aluminum in the alloy substrate. On correlation of the experimentally determined compositions of the final salt films to the calculated phase stability diagrams, the researchers showed that these contaminated salt films were indeed above their liquidus temperatures. Type II hot corrosion was treated in greater detail (Ref 25, 26). Although the details of attack for LTHC differ from those for HTHC, the mechanism for LTCH can also be rationalized in terms of a scale fluxing driven by a negative solubility gradient, as illustrated in Fig. 7.

References cited in this section 5. R.A. Rapp, Corrosion, Vol 42, 1986, p 568 15. R.A. Rapp and K.S. Goto, The Hot Corrosion of Metals by Molten Salts, Molten Salts II, R. Selman and J. Braunstein, Ed., The Electrochemical Society, 1979, p 159 16. M.A. DeCrescente and N.S. Bornstein, Corrosion, Vol 24, 1968, p 127 17. N.S. Bornstein and M.A. DeCrescente, Trans. Metall. Soc. (AIME), Vol 245, 1969, p 1947 18. N.S. Bornstein and M.A. DeCrescente, Corrosion, Vol 26, 1970, p 209 19. N.S. Bornstein and M.A. DeCrescente, Metall. Trans., Vol 2, 1971, p 2875 20. J.A. Goebel and F.S. Pettit, Metall. Trans., Vol 1, 1970, p 1943 21. J.A. Goebel, F.S. Pettit, and G.W. Goward, Metall. Trans., Vol 4, 1973, p 261 22. Y.S. Zhang and R.A. Rapp, Corrosion, Vol 43, 1987, p 348 23. G.C. Fryman, F.J. Kohl, and C.A. Stearns, J. Electrochem. Soc., Vol 131, 1984, p 2985 24. K.L. Luthra and D.A. Shores, J. Electrochem. Soc., Vol 127, 1980, p 2002 25. K.L. Luthra, Metall. Trans. A, Vol 13, 1982, p 1647, 1843, 1853 26. K.L. Luthra, J. Electrochem. Soc., Vol 132, 1985, p 1293


Consequences of Salt Chemistry on Corrosion With the understanding that oxide solubilities in fused salts take on the general behavior exhibited in Fig. 5, that is, at least when simple uncomplexed solutes are involved, a number of concepts helpful to controlling corrosion by fused salts can be understood. In general, the analyses that follow apply to both corrosion in deep melts as well as thin films, although, as stated previously, steeper gradients are expected for hot corrosion involving thin salt films. Negative Solubility Gradient Criterion. Regarding the model for continuing scale dissolution and reprecipitation presented by Fig. 7 and 8, one can inquire whether specific experimental evidence has supported this mechanism. In fact, during the hot corrosion of preoxidized nickel by a fused film of Na 2SO4, researchers (Ref 27) used potentiometric probes to trace the local changes in and log as a function of corrosion time. The experiments were done to respond to a question raised (Ref 28) as to how a NiO scale could be subject to sulfidation and basic fluxing when the typical gas turbine combustion product gases are far to the acid side of the NiO solubility minimum shown in Fig. 5. In equilibrium with an acidic gas, a positive solubility gradient should not lead to NiO dissolution/reprecipitation. The researchers found that for certain (inadequate) preoxidation conditions, the NiO oxide was rapidly attacked by the salt film. Consistent with the model for basic fluxing of Eq 6, the penetration of the salt through the NiO scale to contact the nickel substrate led to a . Simultaneously, sulfidation of the nickel occurred with such a significant increase large decrease in local in the salt film basicity that the negative solubility gradient criterion was indeed satisfied, even for a quite acidic surrounding gas phase. Similar experiments for improved preoxidation conditions (higher nickel purity or thicker scale) showed that the NiO scale was not penetrated, sulfidation and a large shift in basicity did not occur, and the coupon therefore suffered little corrosion attack. These experiments showed the severe consequences of direct salt- metal contact, which causes the local chemistry of the salt film to differ greatly from that in equilibrium with the gas phase. In general, one can conclude that unless the metal is well passivated by a protective oxide, the salt chemistry will be dominated by the reaction at the metal-salt interface. Synergistic Corrosion. On examining the solubility plots for various oxides in Fig. 5, the great displacement in solubilities for the various oxides is apparent. The common base metals for engineering alloys, nickel, cobalt, and iron, form the most basic oxides (subject to acidic dissolution), while the most common protective oxides, chromia and Al2O3, are the most acidic oxides, (subject to basic dissolution). Because the oxidation of an alloy, for example, an iron-chromium alloy (or nickel-chromium, cobalt-chromium, and so on, also involving aluminum) should lead to the presence of both basic and acidic oxides in the protective scale, this situation represents a potential hazard in the presence of a fused salt with a local basicity value between the solubility minima for the two oxides. Therefore, the potential for synergistic dissolution of the protective scale on an ironchromium alloy was studied (Ref 29) according to the following coupled reactions: (Eq 7)

(Eq 8) According to these equations, the product O2- ions for the acidic dissolution of Fe2O3 should serve as a reactant for the basic dissolution of Cr2O3. For fused Na2SO4 salt with a basicity of approximately -13.9 pH (between the relevant minima in Fig. 5), the kinetics of dissolution were measured for powders of Fe 2O3 and Cr2O3, both individually and in the presence of the second oxide. In each case, the mixed oxides dissolved very much faster

than the single oxides. These results show that a mixed oxide protective scale may be subject to accelerated attack by a fused salt. Accordingly, researchers (Ref 24, 25, 26) showed the beneficial effect of high chromium content in nickel-chromium and cobalt-chromium alloys in excluding the oxidation of nickel and cobalt, thereby minimizing type II LTHC. Researchers (Ref 30) showed that cobalt-chromium coatings with more than 37.5% Cr and a small reactive-element addition provide excellent LTHC resistance. Strongly Acidic Oxides. High-temperature engineering alloys frequently contain the alloying elements chromium, tungsten, and molybdenum, while vanadium is often an impurity in low-grade fuels. The oxides of these elements have the tendency to complex with oxide ions in solutions to form chromate, tungstate, molybdate, and vanadate anions. All of these compounds contribute to a reduction in the liquidus temperature in their solutions with sodium sulfate. A lower liquidus temperature means that the potential for hot corrosion by a thin fused salt film is extended to lower temperatures or farther downstream in a combustion system. In addition, the complexing behavior of these oxides introduces an important influence on the acid/base chemistry of the salt and thereby on the scale fluxing mechanism. A phase stability diagram describing the stable vanadate solutes for a sodium sulfate/vanadate solution containing 30 mol% V species at 1173 K (1652 °F) is presented in Fig. 9(a) (Ref 31). Because of their common axes, Fig. 9(a) for the Na-V-S-O system represents a superposition of the Na-V-O phase stabilities onto Fig. 1 for Na2SO4, as was also done in Fig. 3 for the Na-CrS-O system. The fields of dominance for the solutes Na3VO4 (orthovanadate), NaVO3 (metavanadate), and V2O5 (vanadium pentoxide) are apparent in Fig. 9(a), and, under the assumption of an ideal solution, the concentrations of these solutes are plotted in Fig. 9(b) for a sodium sulfate/vanadate solution containing 30 mol% V.

Fig. 9 The Na-V-S-O system. (a) Phase stability diagram at 1173 K (1652 °F). (b) Equilibrium concentrations for Na3VO4, NaVO3, and V2O5 in a sodium sulfate/vanadate solution containing 30 mol% V at 1173 K (1652 °F) The solubilities of the oxides CeO2, HfO2, and Y2O3 were measured as a function of salt basicity at 1173K (1652 °F) and 1 atm O2 (Ref 22). The results were initially very surprising, because the magnitudes for their solubilities were quite high (approximately 100 ppm at their minima), and the solubility curves seem to be shifted toward a relatively basic value. To clarify these apparent discrepancies, the solubility of CeO2 in pure Na2SO4 was measured for comparison. The plot of Fig. 10 shows that the presence of the vanadate component led to a very large increase in the acidic solubility for CeO2, which, accordingly, shifted the solubility minimum to a more basic value (Ref 22). The slopes for the acidic dissolution of CeO2 are in perfect agreement with the reactions: (Eq 9)

(Eq 10)

Because the oxide CeO2 has no special properties, the effect of the strong acid NaVO3, and even more so V2O5, in increasing the acidic solubility and thereby shifting the melt chemistry must be considered a general occurrence expected for all oxides. The magnitudes of these changes would significantly affect the attack of a metal or alloy through the negative solubility gradient mechanism. Other strongly acidic salts would similarly tend to complex with oxide ions to form, for example, molybdate, tungstate, and chromate ions. The salt AlF 3 is also strongly acidic and complexes with oxide ions, for example, in fused chloride or fluoride solutions, to form three different oxyfluoride complexes.

Fig. 10 Measured solubilities of CeO2 in pure Na2SO4 and in 0.7 Na2SO4-0.3NaVO3 at 1173 K (1652 °F) and 1 atm O2 Protective Behavior of Chromium. Finally, one needs to understand why chromium is the most effective alloying element to combat corrosion by sodium sulfate. As illustrated in Fig. 11, the answer seems to lie in the oxygen pressure dependence for the basic dissolution of Cr2O3 (Ref 27). Because the valence of chromium is increased from +3 to +6 on Cr2O3 dissolution to form

ions, the basic solubility of Cr2O3: (Eq 11)

increases with increasing oxygen activity. Figure 11 illustrates the hot corrosion of a Cr2O3-protected alloy threatened by penetration of the salt at some grain boundaries or other defect. Because any thin salt film will be more reducing at the oxide-salt interface (certainly at a metal-salt interface) than farther away from this interface (toward the oxidizing gas), the chromate solute will experience a positive solubility gradient, and

therefore, consumptive reprecipitation of oxide in the salt film will not occur. Rather, the chromate ion will reprecipitate the oxide, satisfying a reduction in its solubility, at the most reduced sites, for example, the flaws or grain boundaries of the scale. In this way, the chromium serves as a protective component to plug flaws in the scale. In aqueous solutions, where chromatic inhibitors have been popular (but are now avoided for environmental reasons), the protective mechanism is believed to be the same.

Fig. 11 The role of chromium in inhibiting hot corrosion attack

References cited in this section 22. Y.S. Zhang and R.A. Rapp, Corrosion, Vol 43, 1987, p 348 24. K.L. Luthra and D.A. Shores, J. Electrochem. Soc., Vol 127, 1980, p 2002 25. K.L. Luthra, Metall. Trans. A, Vol 13, 1982, p 1647, 1843, 1853 26. K.L. Luthra, J. Electrochem. Soc., Vol 132, 1985, p 1293 27. N. Otsuka and R.A. Rapp, J. Electrochem. Soc., Vol 137, 1990, p 46, 53 28. D.A. Shores, in High-Temperature Corrosion NACE-6, R.A. Rapp, Ed., NACE, 1983, p 493 29. Y.S. Hwang and R.A. Rapp, J. Electrochem. Soc., Vol 137, 1990, p 1276 30. K.L. Luthra and J.H. Wood, Thin Solid Films, Vol 119, 1984, p 217 31. Y.S. Hwang and R.A. Rapp, Corrosion, Vol 45, 1989, p 993


Conclusions In this article, the importance of the acid/base and oxidizing character of a fused salt to the corrosion mechanism has been emphasized. In many regards, corrosion by a fused salt shares common characteristics with aqueous corrosion. Corrosion by fused sodium sulfate, also including additions of vanadates, has been emphasized for two reasons: this salt system probably has the greatest practical importance, because it relates to corrosion in combustion product gases; and the fundamental data and models have been better developed for this salt system. For many combinations of fused salt contacting an alloy or other material, the condition leading to failure may be represented by penetration of the salt to contact the reducing substrate, because the salt and the substrate are often thermodynamically incompatible. Consequently, the best protection from fused salt corrosion is then the formation and retention of a dense, adherent protective oxide scale. Because of possible synergistic dissolution, a pure, one-phase protective oxide is preferred. Because of the special solubility dependence for Cr2O3 on oxygen activity, chromium serves as the best protective alloy component to resist attack by an acidic salt. In general, where such knowledge is available, a protective scale should be chosen, such that the minimum in its solubility matches the relevant salt condition. In corrosion testing, care should be given so that the salt and gas compositions and the relevant salt thickness are chosen to match the problem. Provision of an initial protective scale by preoxidation is also useful.


References 1. G.Y. Lai, High-Temperature Corrosion of Engineering Alloys, ASM International, 1990 2. D. Inman and D.M. Wrench, Br. Corros. J., Vol 1, 1966, p 246 3. C.O. Park and R.A. Rapp, J. Electrochem. Soc., Vol 133, 1986, p 1636 4. R. Andresen, J. Electrochem. Soc., Vol 126, 1979, p 328 5. R.A. Rapp, Corrosion, Vol 42, 1986, p 568 6. Y.S. Zhang and R.A. Rapp, J. Met., Vol 46, 1994, p 47 7. Y.S. Zhang, X. Wu, and R.A. Rapp, Metall. Mater. Trans. B, Vol 34, 2003, p 235 8. Y.S. Zhang, J. Electrochem. Soc., Vol 133, 1986, p 655 9. N.S. Jacobson, J. Am. Ceram. Soc., Vol 76, 1993, p 33 10. W.C. Fang and R.A. Rapp, J. Electrochem. Soc., Vol 130, 1983, p 2335 11. Y.M. Wu and R.A. Rapp, J. Electrochem. Soc., Vol 138, 1991, p 2683 12. D.Z. Shi and R.A. Rapp, Werkst. Korros., Vol 41, 1990, p 215

13. J. Kupper and R.A. Rapp, Werkst. Korros., Vol 38, 1987, p 187 14. M. Spiegel, Mater. High Temp., Vol 14, 1998, p 221 15. R.A. Rapp and K.S. Goto, The Hot Corrosion of Metals by Molten Salts, Molten Salts II, R. Selman and J. Braunstein, Ed., The Electrochemical Society, 1979, p 159 16. M.A. DeCrescente and N.S. Bornstein, Corrosion, Vol 24, 1968, p 127 17. N.S. Bornstein and M.A. DeCrescente, Trans. Metall. Soc. (AIME), Vol 245, 1969, p 1947 18. N.S. Bornstein and M.A. DeCrescente, Corrosion, Vol 26, 1970, p 209 19. N.S. Bornstein and M.A. DeCrescente, Metall. Trans., Vol 2, 1971, p 2875 20. J.A. Goebel and F.S. Pettit, Metall. Trans., Vol 1, 1970, p 1943 21. J.A. Goebel, F.S. Pettit, and G.W. Goward, Metall. Trans., Vol 4, 1973, p 261 22. Y.S. Zhang and R.A. Rapp, Corrosion, Vol 43, 1987, p 348 23. G.C. Fryman, F.J. Kohl, and C.A. Stearns, J. Electrochem. Soc., Vol 131, 1984, p 2985 24. K.L. Luthra and D.A. Shores, J. Electrochem. Soc., Vol 127, 1980, p 2002 25. K.L. Luthra, Metall. Trans. A, Vol 13, 1982, p 1647, 1843, 1853 26. K.L. Luthra, J. Electrochem. Soc., Vol 132, 1985, p 1293 27. N. Otsuka and R.A. Rapp, J. Electrochem. Soc., Vol 137, 1990, p 46, 53 28. D.A. Shores, in High-Temperature Corrosion NACE-6, R.A. Rapp, Ed., NACE, 1983, p 493 29. Y.S. Hwang and R.A. Rapp, J. Electrochem. Soc., Vol 137, 1990, p 1276 30. K.L. Luthra and J.H. Wood, Thin Solid Films, Vol 119, 1984, p 217 31. Y.S. Hwang and R.A. Rapp, Corrosion, Vol 45, 1989, p 993

Corrosion by Molten Nitrates, Nitrites, and Fluorides, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 124–128

Corrosion Fluorides

by

Molten

Nitrates,

Nitrites,

and

Introduction MOLTEN SALTS, often called fused salts, are used in many engineering systems. They can cause corrosion by the solution of constituents of the container material, selective attack, pitting, electrochemical reactions, mass transport due to thermal gradients, reaction of constituents of the molten salt with the container material, reaction of impurities in the molten salt with the container material, and reaction of impurities in the molten salt with the alloy. Many hundreds of molten salt/metal corrosion studies have been documented, yet predictions of corrosion are difficult if not impossible. The most prevalent molten salts in use are nitrates and halides. Other molten salts that have been extensively studied but are not widely used include carbonates, sulfates, hydroxides, and oxides.

Corrosion by Molten Nitrates, Nitrites, and Fluorides, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 124–128 Corrosion by Molten Nitrates, Nitrites, and Fluorides

Test Methods A number of kinetic and thermodynamic studies of corrosion by molten salts have been carried out in capsuletype containers. These studies can determine the nature of the corroding species and the corrosion products under static isothermal conditions and do provide some much- needed information. However, to provide the information needed for an actual flowing system, corrosion studies must be conducted in thermal convection loops or forced convection loops, which will include the effects of thermal gradients, flow, chemistry changes, and surface area. These loops can also include electrochemical probes and gas monitors. An example of the types of information gained from thermal convection loops during an intensive study of the corrosion of various alloys by molten salts is given subsequently. A thermal convection loop is shown in Fig. 1.

Fig. 1 Natural circulation loop and salt sampler


Purification Molten salts, whether used for experimental purposes or in actual systems, must be kept free of contaminants. Maintaining purity, which involves initial makeup, transfer, and operation, is specific for each type of molten salt. For example, even though the constituents of the molten fluoride salts used in the Oak Ridge molten salt reactor experiment were available in very pure grades, purification by a hydrogen/hydrogen fluoride (H2/HF) gas purge for 20 h was necessary (Ref 1). For nitrates with a melting point of approximately 220 °C (430 °F), purging with argon flowing above and through the salt at 250 to 300 °C (480 to 570 °F) removes significant amounts of water vapor (Ref 2). Another purification method used for this same type of salt consisted of bubbling pure dry oxygen gas through the 350 °C (660 °F) melt for 2 h and then bubbling pure dry nitrogen for 30 min to remove the oxygen (Ref 3). All metals that contact the molten salt during purification must be carefully selected to avoid contamination from transfer tubes, thermocouple wells, the makeup vessel, and the container itself. This selection process may be an experiment in itself.

References cited in this section 1. J.W. Koger, Report ORNL-TM-4286, Oak Ridge National Laboratory, Dec 1972 2. P.F. Tortorelli and J.H. DeVan, Report ORNL-TM-8298, Oak Ridge National Laboratory, Dec 1982 3. A. Baraka, A.I. Abdel-Rohman, and A.A. El Hosary, Br. Corros. J., Vol 11, 1976, p 44


Nitrates/Nitrites Nitrate mixtures have probably been studied and used more than any other molten salt group. This is perhaps because of the low operating temperatures possible (200 to 400 °C, or 390 to 750 °F). Steels of varying types are generally chosen to contain these systems, because, in general, the basicity of the melt prevents iron corrosion. Protection by passive films is less reliable, because oxide ion discharge may break down the passive film. Electropolished iron spontaneously passivates in molten sodium nitrate/potassium nitrite (NaNO3-KNO2) in the temperature range of 230 to 310 °C (445 to 590 °F) at certain potentials (Ref 4). A magnetite (Fe3O4) film is formed, along with a reduction of nitrite or any trace of oxygen gas dissolved in the melt. At higher potentials, all reactions occur on the passivated iron. Above the passivation potentials, dissolution occurs, with ferric ion dissolving in the melt. At even higher potentials, nitrogen oxides are evolved, and nitrate ions dissolve in the nitrite melt. At higher currents, hematite (Fe2O3) is formed as a suspension, and NO2 is detected. Carbon steel in molten sodium nitrate/potassium nitrate (NaNO3-KNO3) at temperatures ranging from 250 to 450 °C (480 to 840 °F) forms a passivating film consisting mainly of Fe3O4 (Ref 3).

Iron anodes in molten alkali nitrates and nitrites at temperatures ranging from 240 to 320 °C (465 to 610 °F) acquire a passive state in both melts. In nitrate melts, the protective Fe3O4 oxidizes to Fe2O3 and the gaseous products differ for each melt (Ref 5). An interesting study was conducted on the corrosion characteristics of several eutectic molten salt mixtures on such materials as carbon steel, stainless steel, and Inconel in the temperature range of 250 to 400 °C (480 to 750 °F) in a nonflowing system (Ref 6). The salt mixtures and corrosion rates for carbon steel and stainless steel are given in Table 1. As expected, the corrosion rate was much higher for carbon steel than for stainless steel in the same mixture. Low corrosion rates were found for both steels in mixtures containing large amounts of alkaline nitrate. The nitrate ions had a passivating effect. Table 1 Corrosion rates of iron-base alloys in eutectic molten salt mixtures Corrosion rate Carbon steel Stainless steel μm/yr mils/yr μm/yr mils/yr 15 0.6 1 0.03 NaNO3-NaCl-Na2SO4 (86.3,8.4,5.3 mol%, respectively) 23 0.9 7.5 0.3 KNO3-KCl (94.6 mol%, respectively) 63 2.5 20 0.8 LiCl-KCl (58.42 mol%, respectively) Electrochemical studies showed high resistance to corrosion by Inconel. Again, the sulfate- containing mixture caused less corrosion because of the passivating property of the nitrate as well as the preferential adsorption of sulfate ions. Surface analysis by Auger electron spectroscopy indicated varying thicknesses of iron oxide, nickel, and chromium layers. The Auger analysis showed that an annealed and air-cooled stainless steel specimen exposed to molten lithium chloride (LiCl)/potassium chloride (KCl) salt had corrosion to a depth five times greater than that of an unannealed stainless steel specimen. Chromium carbide precipitation developed during slow cooling and was responsible for the increased corrosion. The mechanism of corrosion of iron and steel by these molten eutectic salts can be described by the following reactions: Salt mixture

Fe ↔ Fe2+ + 2e-

(Eq 1)

LiCl + H2O ↔ LiOH + HCl

(Eq 2) (Eq 3)

O2 + 2e- ↔ O2-

(Eq 4)

Fe3+ + e ↔ Fe2+

(Eq 5)

Fe2+ + O2- ↔ FeO

(Eq 6)

3FeO + O2- ↔ Fe3O4 + 2e-

(Eq 7)

2Fe3O4 + O2- ↔ 3Fe2O3 + 2e-

(Eq 8)

Carbon or chromium-molybdenum steels have been used in an actual flowing and operating system of KNO3NaO2-NaO3 (53, 40, and 7 mol%, respectively) at temperatures to 450 °C (840 °F) (Ref 7). For higher temperatures and longer times, nickel or austenitic stainless steels are used. Weld joints are still a problem in both cases. Alloy 800 (UNS N08800) and types 304 (UNS S30400), 304L (UNS S30403), and 316 (UNS S31600) stainless steels were exposed to thermally convective NaNO3-KNO3 salt (draw salt) under argon at 375 to 600 °C (705 to 1110 °F) for more than 4500 h (Ref 2). The exposure resulted in the growth of thin oxide films on all alloys and the dissolution of chromium by the salt. The weight change data for the alloys indicated that the metal in the oxide film constituted most of the metal loss, that the corrosion rate, in general, increased with temperature, and that, although the greatest metal loss corresponded to a penetration rate of 25 μm/yr (1 mil/ yr), the rate was less than 13 μm/yr (0.5 mil/yr) in most cases. These latter rates are somewhat smaller than those reported for similar loops operated with the salt exposed to the atmosphere (Ref 8, 9) but are within a factor of 2 to 5. Spalling had a significant effect on metal loss at intermediate temperatures in the type 304L stainless steel loop.

Metallographic examinations showed no evidence of intergranular attack or of significant cold-leg deposits. Weight change data further confirmed the absence of thermal gradient mass transport processes in these draw salt systems. Raising the maximum temperature of the type 316 stainless steel loop from 595 to 620 °C (1105 to 1150 °F) dramatically increased the corrosion rate (Ref 8, 9). Thus, 600 °C (1110 °F) may be the limiting temperature for use of such alloys in draw salt.

References cited in this section 2. P.F. Tortorelli and J.H. DeVan, Report ORNL-TM-8298, Oak Ridge National Laboratory, Dec 1982 3. A. Baraka, A.I. Abdel-Rohman, and A.A. El Hosary, Br. Corros. J., Vol 11, 1976, p 44 4. A.J. Arvia, J.J. Podesta, and R.C.V. Piatti, Electrochim. Acta, Vol 16, 1971, p 1797 5. A.J. Arvia, J.J. Podesta, and R.C.V. Piatti, Electrochim. Acta, Vol 17, 1972, p 33 6. H.V. Venkatasetty and D.J. Saathoff, International Symposium on Molten Salts, 1976, p 329 7. Yu. I. Sorokin and Kh. L. Tseitlin, Khim. Prom., Vol 41, 1965, p 64 8. R.W. Bradshaw, “Corrosion of 304 Stainless Steel by Molten NaNO3-KNO3 in a Thermal Convection Loop,” SAND-80-8856, Sandia National Laboratory, Dec 1980 9. R.W. Bradshaw, “Thermal Convection Loop Corrosion Tests of 316 Stainless Steel and IN800 in Molten Nitrate Salts,” SAND-81-8210, Sandia National Laboratory, Feb 1982


Fluorides Because of the Oak Ridge molten salt reactor experiment, a large amount of work was done on corrosion by molten fluoride salts (Ref 1). Because these molten salts were to be used as heat- transfer media, temperature gradient mass transfer was very important. Very small amounts of corrosion can result in large deposits, given that the solubility of the corrosion product changes drastically in the temperature range in question. Many other variables can also cause this phenomenon. Thus, a corrosion rate in itself does not provide complete information about corrosion. Because the products of oxidation of metals by fluoride melts are quite soluble in the corroding media, passivation is precluded, and the corrosion rate depends on other factors, including the thermodynamic driving force of the corrosion reactions. The design of a practicable system using molten fluoride salts, therefore, demands the selection of salt constituents, such as lithium fluoride (LiF), beryllium fluoride (BeF 2), uranium tetrafluoride (UF4), and thorium fluoride (ThF4), that are not appreciably reduced by available structural metals and alloys whose components (iron, nickel, and chromium) can be in near-thermodynamic equilibrium with the salt. A continuing program of experimentation over many years has been devoted to defining the thermodynamic properties of many corrosion product species in molten LiF-BeF2 solutions. Much of the data has been obtained by direct measurement of equilibrium pressures for such reactions as:

H2(g) + FeF2(d) ↔ Fe(c) + 2HF(g)

(Eq 9)

and 2HF(g) + BeO(c) ↔ BeF2(l) + H2O(g)

(Eq 10)

where g, c, and d represent gas, crystalline solid, and solute, respectively, using the molten fluoride (denoted 1 for liquid) as the reaction medium. All of these studies have been reviewed, and the combination of these data with those of other studies has yielded tabulated thermodynamic data for many species in molten LiF-BeF 2 (Table 2). From these data, one can assess the extent to which a uranium trifluoride (UF3)-bearing melt will disproportionate according to the reaction: 4UF3(d) ↔ 3UF4(d) + U(d)

(Eq 11)

Table 2 Standard Gibbs free energies (ΔGf) of formation for species in molten 2LiF-BeF2. Temperature range: 733–1000 K Material(a) -ΔGf = c - dT(b) where: -ΔGf at 1000 K, kcal/mol c d 141.8 0.0166 125.2 LiF(l) 243.9 0.0300 106.9 BeF2(l) 338.0 0.0403 99.3 UF3(d) 445.9 0.0579 97.0 UF4(d) 491.2 0.0624 107.2 ThF4(d) 453.0 0.0651 97.0 ZrF4(d) 146.9 0.0363 55.3 NiF2(d) 154.7 0.0218 66.5 FeF2(d) 171.8 0.0214 75.2 CrF2(d) 0.0696 50.2 MoF6(g) 370.9 (a) The standard state for LiF and BeF2 is the molten 2LiF-BeF2 liquid. That for MoF6(g) is the gas at 1 atm. That for all species with d is that hypothetical solution with the solute at unit mole fraction and with the activity coefficient it would have at infinite dilution. (b) Formula gives -ΔGf in kcal/mol for temperature (T) in K in the range 733–1000 K. Source: Ref 10 For the case in which the total uranium content of the salt is 0.9 mol%, as in the Oak Ridge Molten Salt Reactor Experiment, the activity of metallic uranium (referred to as the pure metal) is near 10 -15 with 1% of the UF4 converted to UF3 and is near 2 × 10-10 with 20% of the UF4 so converted (Ref 11). Operation of the reactor with a small fraction (usually 2%) of the uranium present as UF3 is advantageous insofar as corrosion and the consequences of fission are concerned. Such operation with some UF3 present should result in the formation of an extremely dilute (and experimentally undetectable) alloy of uranium with the surface of the container metal. Operation with 50% of the uranium as UF3 would lead to much more concentrated (and highly deleterious) alloying and to formation of uranium carbides. All evidence to date demonstrates that operation with relatively little UF3 is completely satisfactory. The data gathered to date reveal clearly that in reactions with structural metals, M: 2UF4(d) + M(c) ↔ 2UF3(d) + MF2(d)

(Eq 12)

chromium is much more readily attacked than iron, nickel, or molybdenum (Ref 11, 12). Nickel-base alloys, more specifically, Hastelloy N (Ni-6.5Mo-6.9Cr-4.5Fe) and its modifications, are considered the most promising for use in molten salts and have received the most attention. Stainless steels, having more chromium than Hastelloy N, are more susceptible to corrosion by fluoride melts but can be considered for some applications. Oxidation and selective attack may also result from impurities in the melt: M + NiF2 ↔ MF2 + Ni

(Eq 13)

M + 2HF ↔ MF2 + H2

(Eq 14)

or oxide films on the metal:

NiO + BeF2 ↔ NiF2 + BeO

(Eq 15)

followed by reaction of nickel fluoride (NiF2) with M. The reactions given in Eq 13, 14, and 15 will proceed essentially to completion at all temperatures. Accordingly, such reactions can lead (if the system is poorly cleaned) to a rapid initial corrosion rate. However, these reactions do not give a sustained corrosive attack. On the other hand, the reaction involving UF4 (Eq 12) may have an equilibrium constant that is strongly temperature dependent; therefore, when the salt is forced to circulate through a temperature gradient, a possible mechanism exists for mass transfer and continued attack. Equation 12 is of significance mainly in the case of alloys containing relatively large amounts of chromium. If nickel, iron, and molybdenum are assumed to form regular or ideal solid solutions with chromium (as is approximately true), and if the circulation rate is very rapid, the corrosion process for alloys in fluoride salts can be simply described. At high flow rates, uniform concentrations of UF3 and chromium fluoride (CrF2) are maintained throughout the fluid circuit. Under these conditions, there exists some temperature (intermediate between the maximum and minimum temperatures of the circuit) at which the initial chromium concentration of the structural metal is at equilibrium with the fused salt. This temperature, TBP, is called the balance point. Because the equilibrium constant for the chemical reaction with chromium increases with temperature, the chromium concentration in the alloy surface tends to decrease at temperatures higher than TBP and tends to increase at temperatures lower than TBP. At some point, the dissolution process will be controlled by the solidstate diffusion rate of chromium from the matrix to the surface of the alloy. In some melts (NaF-LiF-KF-UF4, for example), the equilibrium constant for Eq 12 with chromium changes sufficiently as a function of temperature to cause the formation of dendritic chromium crystals in the cold zone. For LiF- BeF2-UF4-type mixtures, the temperature dependence of the mass transfer reaction is small, and the equilibrium is satisfied at reactor temperature conditions without the formation of crystalline chromium. Thus, the rate of chromium removal from the salt stream by deposition at cold-fluid regions is controlled by the rate at which chromium diffuses into the cold-fluid wall; the chromium concentration gradient tends to be small, and the resulting corrosion is well within tolerable limits. A schematic of the temperature gradient mass transfer process is shown in Fig. 2.

Fig. 2 Temperature-gradient mass transfer Lithium fluoride/beryllium fluoride salts containing UF4 or ThF4 and tested in thermal convection loops showed temperature gradient mass transfer, as noted by weight losses in the hot leg and weight gains in the cold leg (Fig. 3). Hastelloy N was developed for use in molten fluorides and has proved to be quite compatible. The weight changes of corrosion specimens increased with temperature and time (Fig. 4, 5).

Fig. 3 Weight changes of type 316 stainless steel specimens exposed to LiF-BeF2-ThF4-UF4 (68, 20, 11.7, and 0.3 mol%, respectively) as a function of position (C's, cold leg; H's, hot leg) and temperature

Fig. 4 Weight changes of Hastelloy N specimens versus time of operation in LiF-BeF 2-ThF4 (73, 2, and 25 mol%, respectively)

Fig. 5 Weight changes of Hastelloy N exposed to LiF-BeF2-ThF4-UF4 (68, 20, 11.7, and 0.3 mol%, respectively) for various times A type 304L stainless steel exposed to a fuel salt for 9.5 years (Fig. 6) in a type 304L stainless steel loop showed a maximum uniform corrosion rate of 22 μm/yr (0.86 mil/yr). Voids extended into the matrix for 250 μm (10 mils), and chromium depletion was found (Fig. 7) to extend to a depth of 28 μm (1.1 mil).

Fig. 6 Weight changes of type 304L stainless steel specimens exposed to LiF-BeF2-ZrF4-ThF4-UF4 (70, 23, 5, 1, and 1 mol%, respectively) for various times and temperatures

Fig. 7 Chromium and iron concentration gradient in a type 304L stainless steel specimen exposed to LiFBeF2-ZrF4-ThF4-UF4 (70, 23, 5, 1, and 1 mol%, respectively) for 5700 h at 688 °C (1270 °F) The corrosion resistance of a maraging steel (Fe-12Ni-5Cr-3Mo) at 662 °C (1224 °F) was better than that of type 304L stainless steel but was worse than that of a Hastelloy N under equivalent conditions. As shown in Table 3, the average uniform corrosion rate for the maraging steel was 14 μm/yr (0.55 mil/yr). Voids were seen in the microstructure of the specimens after 5700 h, and electron microprobe analysis disclosed a definite depletion of chromium and iron. Table 3 Comparison of weight losses of alloys at approximately 663 °C (1225 °F) in similar flow fuel salts in a temperature gradient system Weight loss, mg/cm2 Average corrosion 2490 h 3730 h μm/yr mils/yr 3.0 4.8 14 0.55 Maraging steel 10.0 28 1.1 Type 304 stainless steel 6.5 0.4 0.6 1.5 0.06 Hastelloy N Type 316 stainless steel exposed to a fuel salt in a type 316 stainless steel loop showed a maximum uniform corrosion rate of 25 μm/yr (1 mil/ yr) for 4298 h. Mass transfer did occur in the system. For selected nickel- and iron-base alloys, a direct correlation was found between corrosion resistance in molten fluoride salt and chromium and iron content of an alloy. The more chromium and iron in the alloy, the less the corrosion resistance. Alloy

References cited in this section 1. J.W. Koger, Report ORNL-TM-4286, Oak Ridge National Laboratory, Dec 1972 10. C.F. Baes, Jr., “The Chemistry and Thermodynamics of Molten Salt Reactor Fuels,” Paper presented at the AIME Nuclear Fuel Reprocessing Symposium, (Ames, IA), American Institute of Mining, Metallurgical, and Petroleum Engineers, Aug 1969; see also 1969 Nuclear Metallurgy Symposium, Vol 15, United States Atomic Energy Commission Division of Technical Information Extension 11. G. Long, “Reactor Chemical Division Annual Program Report,” ORNL-3789, Oak Ridge National Laboratory, Jan 1965, p 65

12. J.W. Koger, “MSR Program Semiannual Progress Report,” ORNL-4622, Oak Ridge National Laboratory, Aug 1970, p 170


Conclusions In order to study the corrosion of molten salts or to determine what materials are compatible with a certain molten salt, the following questions must be answered. What is the purpose of the investigation? Is the researcher interested in basic studies, or is this work for information or work preliminary to assessment for a real system? For basic studies, capsule experiments or information from capsules is sufficient. Otherwise, flow systems or information from flow systems will be needed at some point to assess temperature gradient mass transfer. Salts to be used in either case need to be purified, and the same purity must be used in each experiment, unless this factor is a variable. Analytical facilities must be used for the chemistry of the salt, including impurity content and surface analysis of the metals in question. Useful information can be obtained from capsule and flow experiments. It is hoped that the preceding information on specific systems provides an appreciation of the challenges involved and the materials information that can be obtained from various experiments.


Acknowledgment Portions of this article were adapted from J.W. Koger, Fundamentals of High-Temperature Corrosion in Molten Salts, Corrosion, Vol 13, ASM Handbook (formerly Metals Handbook, 9th ed.), ASM International, 1987, p 50–55.


References 1. J.W. Koger, Report ORNL-TM-4286, Oak Ridge National Laboratory, Dec 1972 2. P.F. Tortorelli and J.H. DeVan, Report ORNL-TM-8298, Oak Ridge National Laboratory, Dec 1982

3. A. Baraka, A.I. Abdel-Rohman, and A.A. El Hosary, Br. Corros. J., Vol 11, 1976, p 44 4. A.J. Arvia, J.J. Podesta, and R.C.V. Piatti, Electrochim. Acta, Vol 16, 1971, p 1797 5. A.J. Arvia, J.J. Podesta, and R.C.V. Piatti, Electrochim. Acta, Vol 17, 1972, p 33 6. H.V. Venkatasetty and D.J. Saathoff, International Symposium on Molten Salts, 1976, p 329 7. Yu. I. Sorokin and Kh. L. Tseitlin, Khim. Prom., Vol 41, 1965, p 64 8. R.W. Bradshaw, “Corrosion of 304 Stainless Steel by Molten NaNO3-KNO3 in a Thermal Convection Loop,” SAND-80-8856, Sandia National Laboratory, Dec 1980 9. R.W. Bradshaw, “Thermal Convection Loop Corrosion Tests of 316 Stainless Steel and IN800 in Molten Nitrate Salts,” SAND-81-8210, Sandia National Laboratory, Feb 1982 10. C.F. Baes, Jr., “The Chemistry and Thermodynamics of Molten Salt Reactor Fuels,” Paper presented at the AIME Nuclear Fuel Reprocessing Symposium, (Ames, IA), American Institute of Mining, Metallurgical, and Petroleum Engineers, Aug 1969; see also 1969 Nuclear Metallurgy Symposium, Vol 15, United States Atomic Energy Commission Division of Technical Information Extension 11. G. Long, “Reactor Chemical Division Annual Program Report,” ORNL-3789, Oak Ridge National Laboratory, Jan 1965, p 65 12. J.W. Koger, “MSR Program Semiannual Progress Report,” ORNL-4622, Oak Ridge National Laboratory, Aug 1970, p 170



C.B. Allen and G.T. Janz, J. Hazard. Mater., Vol 4, 1980, p 145 R.W. Bradshaw and S.H. Goods, Corrosion Resistance of Stainless Steels during Thermal Cycling in Alkali Nitrate Molten Salts, “Sandia Report, SAND2001-8518,” Sept. 2001, 36 p R.J. Gale and D.G. Lovering, in Molten Salt Techniques, Vol 2, Plenum Press, 1984, p 1 D. Inman and D.G. Lovering, in Comprehensive Treatise of Electrochemistry, Vol 7, Plenum Publishing, 1983 G.J. Janz and R.P.T. Tompkins, in Corrosion, Vol 35, NACE International, 1979, p 485 C.A.C. Sequeira, Electrochemistry of Corrosion in Molten Salts, Trans Tech Publications, Inc., 2003 C.A.C. Sequeira, High Temperature Corrosion in Molten Salts, Molten Salt Forum, Vol 7-7, Trans Tech Publications, Inc., 2003 Tz. Tzvetkoff, Mechanism of Growth, Composition and Structure of Passive Films Formed and Their Alloys in Molten Salt Electrolytes, Trans Tech Publications, Inc., 2003

P.F. Tortorelli and S.J. Pawel, Corrosion by Liquid Metals, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 129–134

Corrosion by Liquid Metals P.F. Tortorelli and S.J. Pawel, Oak Ridge National Laboratory

Introduction CONCERN ABOUT CORROSION of solids exposed to liquid metal environments, that is, liquid metal corrosion, dates from the earliest days of metals processing, when it became necessary to handle and contain molten metals. Corrosion considerations also arise when liquid metals are used in applications that exploit their chemical or physical properties. Liquid metals serve as high-temperature reducing agents in the production of metals (such as the use of molten magnesium to produce titanium), and, because of their excellent heat- transfer properties, liquid metals have been used or considered as coolants in a variety of power- producing systems. Examples of such applications include molten sodium for liquid metal fast breeder reactors and central receiver solar stations as well as liquid lithium for fusion and space nuclear reactors. In addition, tritium breeding in deuterium-tritium fusion reactors necessitates the exposure of lithium atoms to fusion neutrons. Breeding fluids of lithium or lead-lithium are attractive for this purpose. Molten lead or bismuth can serve as neutron multipliers to raise the tritium breeding yield if other types of lithium-containing breeding materials are used. More recently, in the United States, liquid mercury has been selected as the target for the spallation neutron source (SNS) at Oak Ridge National Laboratory. In the SNS scheme, 1 to 2 GeV protons will impinge on flowing mercury to generate an intense source of spallation neutrons. (Several other planned neutron sources worldwide may also use mercury for this application.) Liquid metals can also be used as two-phase working fluids in Rankine cycle power conversion devices (molten cesium or potassium) and in heat pipes (potassium, lithium, sodium, sodium-potassium). Because of their high thermal conductivities, sodium-potassium alloys, which can be any of a wide range of sodium-potassium combinations that are molten at or near room temperature, have also been used as static heat sinks in automotive and aircraft valves. Whenever the handling of liquid metals is required, whether in specific uses as discussed previously or as melts during processing, a compatible containment material must be selected. At low temperatures, liquid metal corrosion is often insignificant, but in more demanding applications, corrosion considerations can be important in selecting the appropriate containment material and operating parameters. Thus, liquid metal corrosion studies in support of heat pipe technology and aircraft, space, and fast breeder reactor programs date back many years and, more recently, are being conducted worldwide as part of the fusion energy and spallation neutron programs. In this article, the principal corrosion reactions and important parameters that control such processes are briefly reviewed for materials (principally metals) exposed to liquid metal environments. Only corrosion phenomena are covered, and the discussion is limited to corrosion under single-phase (liquid) conditions.

P.F. Tortorelli and S.J. Pawel, Corrosion by Liquid Metals, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 129–134 Corrosion by Liquid Metals P.F. Tortorelli and S.J. Pawel, Oak Ridge National Laboratory

Corrosion Reactions in Liquid Metal Environments

Liquid metal corrosion can manifest itself in various ways. In the most general sense, the following categories can be used to classify relevant corrosion phenomena: • • • •

Dissolution Impurity and interstitial reactions Alloying Compound reduction

Definitions and descriptions of these types of reactions are given subsequently. However, it is important to note that this classification is somewhat arbitrary, and, as will become clear during the following discussion, the individual categories are not necessarily independent of one another. All of these reactions require the wetting of the solid by the liquid metal. This can be inherently problematic in certain systems or in cases when a solid that would normally be wet by the liquid metal has a surface film that prevents such wetting.

Dissolution The simplest corrosion reaction that can occur in a liquid metal environment is direct dissolution. Direct dissolution is the release of atoms of the containment material into the melt in the absence of any impurity effects. Such a reaction is a simple solution process and therefore is governed by the elemental solubilities in the liquid metal and the kinetics of the rate-controlling step of the dissolution reaction. The net rate, J, at which an elemental species enters solution, can be described as: J = k(C0 - C)

(Eq 1)

where J is the rate of mass loss of an element due to dissolution, C0 is the solubility of that element in the liquid metal, C is its actual concentration in the bulk of the liquid, and k is the effective solution rate constant. Each constituent of the condensed phase is separately governed by Eq 1. Because J is proportional to (C0 - C), high solubilities do not necessarily translate into high dissolution rates, because the driving force for dissolution depends on the actual concentration of the solute in the liquid metal as well as C0. This forms the crucial differentiation between isothermal liquid metal systems and nonisothermal dynamic ones regarding the stability of a material with respect to dissolution and mass transfer (see the following discusion). Under isothermal conditions, the rate of this dissolution reaction would decrease with time as C increases. After a period of time, the actual elemental concentration becomes equal to the solubility, and the dissolution rate is then 0. Therefore, in view of Eq 1, corrosion by the direct dissolution process can be minimized by selecting a containment material whose elements have low solubilities in the liquid metal of interest and/or by saturating the melt before actual exposure. However, if the dissolution kinetics are relatively slow, that is, for low values of k, corrosion may be acceptable for short-term exposures. The functional dependence and magnitude of the solution rate constant, k, depend on the rate-controlling step, which, in the simplest cases, can be a transport across the liquid-phase boundary layer, diffusion in the solid, or a reaction at the phase boundary. Measurements of mass changes as a function of time for a fixed C0 - C (see subsequent discussion) yield the kinetic information necessary for determination of the rate-controlling mechanism. Corrosion resulting from dissolution in a nonisothermal liquid metal system is more complicated than the isothermal case. Both solvation and solute deposition are normally occurring, and there is essentially a timeinvariant (C0 - C) driving force for dissolution at a given temperature. This is because, after a short transient period, C is constant as the flow of the liquid metal quickly establishes a systemwide steady state. Specifically, the value of C is set by the dissolution and deposition rates, the temperature profile, and the functional dependence of C0 on temperature. As such, C does not vary around a flowing liquid metal system and, at steady state, is constant with time. Unlike the isothermal case, J is only dependent on time through its dependence on k (Eq 1). The flux description illustrated by Eq 1 applies equally well to deposition when C > C0. Therefore, at temperatures where the solubility (C0) is greater than the bulk concentration (C), dissolution of an element into the liquid metal will occur, but at lower temperatures in the circuit where C0 < C, a particular element will tend to come out of solution and be deposited on the containment material (or remain as suspended particles). A schematic of such a mass-transfer process is shown in Fig. 1. Such mass-transfer processes under nonisothermal conditions can be of prime importance when, in the absence of dissimilar-metal effects (see subsequent

discussion), forced circulation (pumping) of liquid metals used as heat-transfer media exacerbates the transport of materials from hotter to cooler parts of the liquid metal circuit.

Fig. 1 Schematic of thermal gradient mass transfer in a liquid metal circuit. Source: Ref 1 If the dissolution and deposition behaviors are controlled by the respective forms of Eq 1, measurements of mass change as a function of loop position (x) should yield data resembling the schematic depiction shown in Fig. 2. Using the appropriate analysis methods and simplifying assumptions, measurements of the mass or, for alloys/compounds, composition change as a function of loop position (and, therefore temperature) can provide further data regarding kinetics and the rate-controlling mechanism of mass transport associated with a given material/liquid metal system. This can be computed for any element in a closed tubular liquid metal system where deposition occurs over an area Ap and dissolution over As, using the following equation: G = ∫AsJs · 2πrdx - ∫ApJp · 2πrdx

(Eq 2)

where G is the net increase of an element in the circuit per unit time, r is the inside tube radius, Js is as defined in Eq 1 for dissolution, and Jp is of the same functional form but applies to deposition of an element from the liquid metal with a rate constant kp. Typically, G ≈ 0 at steady state, so, using Eq 1 with Eq 2: (Eq 3)

Fig. 2 Schematic depiction of idealized profile of mass change per unit surface area for a liquid metal circuit. Tmax and Tmin are maximum and minimum loop temperatures, respectively; x is loop position; Ap is deposition area; and As is dissolution area. The respective areas of dissolution and deposition, As and Ap, are delimited by two balance points, which are the values of x at which J = 0 (Fig. 2). These balance points are experimentally accessible using an appropriately configured nonisothermal liquid metal loop and are important in determining the relative areas of dissolution and deposition and in modeling the relevant mass-transfer processes. Assuming a linear dependence of C0 over the relatively small temperature range of most loop experiments, the mass balance represented by Eq 3 indicates that the ratio of the area measurements for dissolution and deposition is inversely proportional to the ratio of their respective rate constants: As/Ap α kp/ks

(Eq 4)

Thus, determination of mass change or surface composition profiles around a loop yields insight into the kinetics of the solution process relative to deposition for a given material/liquid metal system. Another use of mass or compositional profiles is to determine estimates of the activation energy for dissolution using reasonable assumptions about the temperature dependence of the solubility. Because the respective activation energies characteristic of the most important rate-controlling processes (surface reaction, liquid-phase diffusion through boundary layer, and solid-state diffusion) are significantly different from each other, such analyses can yield mechanistic information regarding dissolution processes. Although the mass-transfer analysis described previously is somewhat simplified, experimental measurements of mass or composition changes around a nonisothermal liquid metal loop often yield results that can be described in this way. Figure 3 is an example of an experimentally determined mass-change profile consistent with the schematic one shown in Fig. 2. Sometimes, however, more elaborate analyses based on Eq 1 are required to describe nonisothermal mass transfer precisely. Such treatments must take into account the differences in k around the circuit as well as the possibility that the rate constant for dissolution (or deposition) may not vary monotonically with temperature because of changes in the rate-controlling step within the temperature range of dissolution (deposition). The presence of more than one elemental species in the containment material further complicates the analysis; the transfer of each element typically has to be handled with its own set of thermodynamic and kinetic parameters. Although a thermal gradient increases the amount of dissolution, plugging of coolant pipes by nonuniform deposition of dissolved species in cold zones often represents a more serious design problem than metal loss from dissolution (which sometimes may be handled by corrosion allowances). The most direct way to control deposition, however, is usually to minimize dissolution in the hot zone by use of more corrosion-resistant materials and/or inhibition techniques.

Fig. 3 Mass transfer as characterized by the weight changes of type 316 stainless steel coupons exposed around a nonisothermal liquid lithium type 316 stainless steel circuit for 9000 h. Source: Ref 2 Mass transfer may occur even under isothermal conditions if an activity gradient based on composition differences of solids in contact with the same liquid metal exists in the system. Under the appropriate conditions, dissolution and deposition will act to equilibrate the activities of the various elements in contact with the liquid metal. Normally, such a process is chiefly limited to interstitial element transfer between dissimilar metals, but transport of substitutional elements can also occur. Elimination (or avoidance) of concentration (activity) gradients across a liquid metal system is the obvious and, most often, the simplest solution to any problems arising from this type of mass-transport process. Under certain conditions, dissolution of metallic alloys by liquid metals can lead to irregular attack (Fig. 4). Although such localized corrosive attack can often be linked to impurity effects (see subsequent discussion) and/or compositional inhomogeneities in the solid, destabilization of a planar surface can occur when there is preferential dissolution of one or more elements of an alloy exposed to a liquid metal. Indeed, the type of attack illustrated in Fig. 4 is thought to be caused by the preferential dissolution of nickel from an Fe-17Cr-11Ni (wt%) alloy (type 316 stainless steel). As such, this process resembles the dealloying phenomenon sometimes observed in aqueous environments. This type of attack has often been observed for austenitic stainless steels in molten lithium, sodium, lead, and other liquid metals when preferential dissolution of nickel occurs. In contrast, an Fe-12Cr- 1MoVW steel did not experience preferential loss of any of its elements and corroded uniformly (Fig. 5) when exposed under the same environment conditions that led to the irregular attack shown in Fig. 4. In cases where preferential dissolution and nonuniform corrosion occur (Fig. 4), the altered near-surface zone has diminished load-bearing capacity, and the section thickness of sound material remaining will be less than what would be calculated based on converting measured mass losses to a surface recession distance.

Fig. 4 Polished cross section of type 316 stainless steel exposed to thermally convective Pb-17at.%Li at 500 °C (930 °F) for 2472 h. Source: Ref 3

Fig. 5 Polished cross section of Fe-12Cr-1MoVW steel exposed to thermally convective Pb-17at.%Li at 500 °C (930 °F) for 2000 h. Source: Ref 3 Apart from possible effects on morphological development, the changes in surface composition due to preferential elemental dissolution from an alloy into a liquid metal are important in themselves. For example, in austenitic stainless steels, the preferential dissolution of nickel causes a phase transformation to a ferritic structure in the surface region. In many cases, an equilibrium surface composition is achieved, such that the net elemental fluxes into the liquid metal are in the same proportion as the starting concentrations of these elements in the alloy. Such a phenomenon has been rigorously treated and characterized for sodium-steel systems.

Impurity and Interstitial Reactions For this discussion, impurity or interstitial reactions refer to the interaction of light elements present in the containment material (interstitials) or the liquid metal (impurities). Examples of such reactions include the decarburization of steel in lithium and the oxidation of steel in sodium or lead of high oxygen activity. In many cases, when the principal elements of the containment material have low solubilities in liquid metals (for

example, refractory metals in sodium, lithium, and lead), reactions involving light elements such as oxygen, carbon, and nitrogen dominate the corrosion process. Impurity or interstitial reactions can be generally classified into two types: corrosion product formation and elemental transfer of such species. Corrosion Product Formation. The general form of a corrosion product reaction is: xL + yM + zI = LxMyIz

(Eq 5)

where L is the chemical symbol for a liquid metal atom, M is one species of the containment material, and I represents an interstitial or impurity atom in the solid or liquid (x, y, z > 0). The LxMyIz corrosion product that forms by such a reaction may be soluble or insoluble in the liquid metal. If it is soluble, the I species would cause greater dissolution weight losses and would result in an apparently higher solubility of M in L (Eq 1). This is a frequent cause of erroneous solubility measurements and is a good illustration of how dissolution and impurity reactions can be interrelated. Furthermore, if a soluble corrosion product forms at selected sites on the surface of the solid, localized attack will result. Under conditions in which a corrosion product is insoluble, a partial or complete surface layer will form. This can either have a detrimental effect on corrosion resistance or, if the product is continuous and protective (see subsequent discussion), limit further reaction. However, formation of a surface product does not necessarily mean that it can be observed. The product may be unstable outside the liquid metal environment or may dissolve in the cleaning agent used to remove the solidified residue of liquid metal from the exposed containment material. A good example of the importance of impurity or interstitial reactions that form corrosion products can be found in the sodium-steel-oxygen system. It is thought that the reaction: (Eq 6) increases the apparent solubility of iron in sodium at higher oxygen activities, while the interaction of oxygen, sodium, and chromium can lead to the formation of surface corrosion products, for example: 2Na2O(l) + Cr(s) = NaCrO2(s) + 3Na(l)

(Eq 7)

This second type of reaction (Eq 7) is of primary importance in the corrosion of chromium-containing steels by liquid sodium. It can be controlled by reducing the oxygen concentration of the sodium to less than approximately 3 ppm and/or by modifying the composition of the alloy through reduction of the chromium concentration of the steel. Such corrosion product reactions can also be observed in lithium-steel systems, in which nitrogen can increase the corrosiveness of the liquid metal environment. In particular, the reaction: 5Li3N(in l) + Cr(s) = Li9CrN5(s) + 6Li(l)

(Eq 8)

or an equivalent one with iron, can play an important role in corrosion by liquid lithium. The Li9CrN5 corrosion product tends to be localized at the grain boundaries of exposed steels. Such reaction products can probably also be formed when there is sufficient nitrogen in the solid; experimental observations have indicated that nitrogen can increase corrosion by lithium, whether it is in the liquid metal or in the steel. Corrosion product formation is also important when certain refractory metals are exposed to molten lithium. Despite their low solubilities in lithium, niobium and tantalum can be severely attacked when exposed to lithium if the oxygen activities of these metals are not low. At temperatures below approximately 900 °C (1650 °F), the lithium reacts rapidly with the oxygen and niobium or tantalum (and their oxides and suboxides) to form a ternary oxide corrosion product. Such reactions result in localized penetration along grain boundaries and selected crystallographic planes. This form of corrosive attack can be eliminated, however, by minimizing the oxygen concentration of these refractory metals (Fig. 6) and by using alloying additions that form oxides that do not react with the lithium and that minimize the amount of uncombined oxygen in the material (1 to 2 at. % Zr in niobium and hafnium in tantalum).

Fig. 6 Effect on initial oxygen concentration (150 to 1700 ppm) in niobium on the depth of reaction with lithium to form a corrosion product. Polished and etched cross sections of niobium exposed to isothermal lithium at 816 °C (1500 °F) for 100 h. (a) 150 ppm. (b) 500 ppm. (c) 1000 ppm. (d) 1700 ppm. Etched with HF-HNO3-H2SO4-H2O. Source: Ref 4 A final example of a corrosion product reaction that can occur in a liquid metal environment is the oxidation of a solid metal or alloy exposed to molten lead somewhat enriched in oxygen. In some cases, this reaction may actually be beneficial by providing a protective barrier against the highly aggressive lead. This barrier can act in a manner analogous to the behavior observed for the protective oxides formed in high-temperature oxidizing gases. However, this surface product will form and then heal only when the oxygen activity of the melt is maintained at a high level or when oxide formers, such as aluminum or silicon, have been added to the containment alloy to promote protection by the formation of alumina- or silica-containing surface products. Furthermore, reactions of additives to the melt with nitrogen in steel to form nitride surface films are thought to be the cause of reduced corrosion in lead and lead-bismuth systems. Elemental Transfer of Impurities and Interstitials. The second general type of impurity or interstitial reaction is that of elemental transfer. In contrast to what is defined as corrosion product formation, elemental transfer manifests itself as a net transfer of interstitials or impurities to, from, or across a liquid metal. Although compounds may form or dissolve as a result of such transfer, the liquid metal atoms do not participate in the formation of stable products by reaction with the containment material. For example, because lithium is such a

strong thermodynamic sink for oxygen, exposure of oxygen-containing metals and alloys to this liquid often results in the transfer of oxygen to the melt. Indeed, for oxygen-contaminated niobium and tantalum, hightemperature lithium exposures result in the rapid movement of oxygen into the lithium. The thermodynamic driving force for light element transfer between solid and liquid metals is normally expressed in terms of a distribution (or partitioning) coefficient. This distribution coefficient is the equilibrium ratio of the concentration of an element, such as oxygen, nitrogen, carbon, or hydrogen, in the solid metal or alloy to that in the liquid. Such coefficients can be calculated from knowledge or estimates of free energies of formation and activities based on equilibrium between a species in the solid and liquid. An example of this approach is its application to decarburization/carburization phenomena in a liquid metal environment. Carbon transfer to or from the liquid metal can cause decarburization of iron-chromium-molybdenum steels, particularly lower-chromium steels, and carburization of refractory metals and higher-chromium alloys. There have been many studies of such reactions for sodium-steel systems. Although less work has been done in the area of lithium-steel carbon transfer, the same considerations apply. Specifically, the equilibrium partitioning of the carbon between the iron-chromium-molybdenum steel and the lithium can be described as: (Eq 9) where Cc(s), Cc(Li) is the concentration of carbon in the steel and lithium, respectively; aCr is the chromium activity of the steel; C°C(s), C°C(Li) represent the equilibrium solubilities of carbon in the steel and lithium, respectively; x, y is the stoichiometry of the chromium carbide; and:

where ΔF represents the free energies of formation of the indicated compounds, R is the gas constant, and T is the absolute temperature. Equation 9 indicates that in order to decrease the tendency for decarburization of an alloy— that is, to increase the partitioning coefficient, Cc(s)/Cc(Li)—the chromium activity of the alloy must be increased or a more thermodynamically stable carbide dispersion must be developed (by alloy manipulation or thermal treatment). Experiments in lithium and sodium have shown that these factors have the desired effect. Tempering of ironchromium-molybdenum steels to yield more stable starting carbides can significantly reduce decarburization by these two liquid metals. With very unstable microstructures, the steel can be severely corroded because of rapid lithium attack of the existing carbides. Furthermore, alloying additions, such as niobium, form very stable carbides and can dramatically reduce decarburization. In addition, as shown by Eq 9, increasing the chromium level of a steel effectively decreases the tendency for carbon loss. With higher-chromium steels, for example, austenitic stainless steels, carburization can then become a problem. If two dissimilar steels of significantly differing chromium activities and/or microstructures are exposed to the same liquid metal, the melt can act as a conduit for the relatively rapid redistribution of carbon between the two solids. Similar considerations would apply for any light element transfer across a liquid metal in contact with dissimilar materials; this can be further complicated by concentration (activity) gradient mass transfer of substitutional elements, as discussed previously.

Alloying Reactions between atoms of the liquid metal and those of the constituents of the containment material may lead to the formation of a stable product on the solid without the participation of impurity or interstitial elements: xM + yL = MxLy

(Eq 10)

This is not a common form of liquid metal corrosion, particularly with the molten alkali metals, but it can lead to detrimental consequences if it is not understood or anticipated. Alloying reactions, however, can be used to inhibit corrosion by adding an element to the liquid metal to form a corrosion-resistant layer by reaction of this species with the contaminant material. An example is the addition of aluminum to a lithium melt contained by steel. A more dissolution-resistant aluminide surface layer forms, and corrosion is reduced.

Compound Reduction Attack of ceramics exposed to liquid metals can occur because of reduction of the solid by the melt. In very aggressive situations, such as when most oxides are exposed to molten lithium, the effective result of such exposure is the loss of structural integrity by reduction-induced removal of the nonmetallic element from the solid. The tendency for reaction under such conditions can be qualitatively evaluated by consideration of the free energy of formation of the solid oxide relative to the oxygen/oxide stability in the liquid metal. Similar considerations apply to the evaluation of potential reactions between other nonmetallic compounds (nitrides, carbides, and so on) and liquid metals.

References cited in this section 1. J.E. Selle and D.L. Olson, in Materials Considerations in Liquid Metal Systems in Power Generation, National Association of Corrosion Engineers, 1978, p 15–22 2. P.F. Tortorelli and J.H. DeVan, J. Nucl. Mater., Vol 85 and 86, 1979, p 289–293 3. P.F. Tortorelli and J.H. DeVan, J. Nucl. Mater., Vol 141–143, 1986, p 592–598 4. J.R. DiStefano and E.E. Hoffman, Corrosion Mechanisms in Refractory Metal-Alkali Metal Systems, At. Energy Rev., Vol 2, 1964, p 3–33


Considerations in Materials Selection The previously mentioned types of corrosion reactions must be considered in materials selection for liquid metal containment. In many cases, particularly at low temperature or with less aggressive liquids (such as molten steel), liquid metal corrosion is not an important factor, and many materials, both metals and ceramics, would suffice. Under more severe conditions, however, an understanding of the various types of liquid metal corrosion is necessary to select or develop a compatible containment material. For example, for applications in high-temperature molten lithium, most oxides would be unstable with respect to this liquid metal, lowchromium steels would decarburize, and alloys containing large amounts of nickel or manganese would suffer extensive preferential dissolution and irregular attack. Materials selection would then be limited to higherchromium ferritic/martensitic steels or high-purity refractory metals and alloys. A general summary of the types of the most common corrosion reactions and guidelines for materials selection and/or development is given in Table 1, which also includes typical examples for each category. Because two or more concurrent corrosion reactions are possible, and because consideration of all of the applicable materials consequences may lead to opposite strategies, materials selection for liquid metal environments can become quite complex and may require optimization of several factors rather than minimization of any particular one. In addition, an assessment of the suitability of a given material for liquid metal service must be based on the knowledge of its total corrosion response. As in many corrosive environments, a simple numerical rate is not an accurate measurement of the susceptibility of a material when reaction with the liquid metal results in more than one of the modes of attack shown in Fig. 7 and discussed previously. Under such circumstances, a

measurement reflecting total corrosion damage is much more appropriate for judging the ability of a material to resist corrosion by a particular liquid metal. Table 1 Guidelines for materials selection and/or alloy development based on liquid metal corrosion reactions Corrosion reaction Direct dissolution

Guidelines Lower activity of key elements.

Example Reduce nickel for lithium, lead, or sodium systems. Corrosion product Lower activity of reacting elements. Reduce chromium and nitrogen in lithium formation systems. In case of protective oxide, add elements Add aluminum or silicon to steel exposed to promote formation. to lead. Elemental transfer Increase (or add) elements to decrease Increase chromium content in steels transfer tendency. exposed to sodium or lithium. Minimize element being transferred. Reduce oxygen content in metals exposed to lithium. Alloying Avoid systems that form stable Do not expose nickel to molten aluminum. compounds. Promote formation of corrosion-resistant Add aluminum to lithium to form surface layers by alloying. aluminides. Compound reduction Eliminate solids that can be reduced by Avoid bulk oxide-lithium couples. liquid metal.

Fig. 7 Representative modes of surface damage in liquid metal environments. IGA, intergranular attack. Source: Ref 5

Reference cited in this section 5. J.H. DeVan and C. Bagnall, in Proceedings of the International Conference on Liquid Metal Engineering and Technology, Vol 3, The British Nuclear Energy Society, 1985, p 65–72


Acknowledgment Research sponsored by the U.S. Department of Energy under contract DE-AC05-00OR22725 with UT-Battelle, LLC.


References 1. J.E. Selle and D.L. Olson, in Materials Considerations in Liquid Metal Systems in Power Generation, National Association of Corrosion Engineers, 1978, p 15–22 2. P.F. Tortorelli and J.H. DeVan, J. Nucl. Mater., Vol 85 and 86, 1979, p 289–293 3. P.F. Tortorelli and J.H. DeVan, J. Nucl. Mater., Vol 141–143, 1986, p 592–598 4. J.R. DiStefano and E.E. Hoffman, Corrosion Mechanisms in Refractory Metal-Alkali Metal Systems, At. Energy Rev., Vol 2, 1964, p 3–33 5. J.H. DeVan and C. Bagnall, in Proceedings of the International Conference on Liquid Metal Engineering and Technology, Vol 3, The British Nuclear Energy Society, 1985, p 65–72


Selected References • • •

•

T.L. Anderson and G.R. Edwards, The Corrosion Susceptibility of 2 Cr-1 Mo Steel in a Lithium-17.6 Wt Pct Lead Liquid, J. Mater. Energy Syst., Vol 2, 1981, p 16–25 R.C. Asher, D. Davis, and S.A. Beetham, Some Observations on the Compatibility of Structural Materials with Molten Lead, Corros. Sci., Vol 17, 1977, p 545–547 M.G. Barker, S.A. Frankham, and N.J. Moon, The Reactivity of Dissolved Carbon and Nitrogen in Liquid Lithium, Proceedings of the Third International Conference on Liquid Metal Engineering and Technology, Vol 2, The British Nuclear Energy Society, 1984, p 77–83 M.G. Barker, P. Hubberstey, A.T. Dadd, and S.A. Frankham, The Interaction of Chromium with Nitrogen Dissolved in Liquid Lithium, J. Nucl. Mater., Vol 114, 1983, p 143–149

•

•

•

• • •

• • • •

•

• • • • • • • • • •

• •

G.E. Bell, M.A. Abdou, and P.F. Tortorelli, Experimental and Analytical Investigations of Mass Transport Processes of 12Cr- 1MoVW Steel in Thermally Convected Lithium Systems, Fusion Eng. Des., Vol 8, 1989, p 421–427 N.M. Beskorovainyi, V.K. Ivanov, and M.T. Zuev, Behavior of Carbon in Systems of the Metal-Molten Lithium-Carbon Type, High- Purity Metals and Alloys, V.S. Emel'yanov and A.I. Evstyukhin, Ed., Consultants Bureau, 1967, p 107–119 N.M. Beskorovainyi and V.K. Ivanov, Mechanism Underlying the Corrosion of Carbon Steels in Lithium, High-Purity Metals and Alloys, V.S. Emel'yanov and A.I. Evstyukhin, Ed., Consultants Bureau, 1967, p 120–129 O.K. Chopra, K. Natesan, and T.F. Kassner, Carbon and Nitrogen Transfer in Fe-9Cr-Mo Ferritic Steels Exposed to a Sodium Environment, J. Nucl. Mater., Vol 96, 1981, p 269–284 O.K. Chopra and P.F. Tortorelli, Compatibility of Materials for Use in Liquid-Metal Blankets of Fusion Reactors, J. Nucl. Mater., Vol 122 and 123, 1984, p 1201–1212 L.F. Epstein, Static and Dynamic Corrosion and Mass Transfer in Liquid Metal Systems, Liquid Metals Technology—Part I, Vol 53 (No. 20), Chemical Engineering Progress Symposium Series, American Institute of Chemical Engineers, 1957, p 67–81 T. Flament, P.F. Tortorelli, V. Coen, and H.U. Borgstedt, Compatibility of Materials in Fusion First Wall and Blanket Structures Cooled by Liquid Metals, J. Nucl. Mater., Vol 191–194, 1992, p 132–138 J.D. Harrison and C. Wagner, The Attack of Solid Alloys by Liquid Metals and Salt Melts, Acta Metall., Vol 1, 1959, p 722–735 E.E. Hoffman, “Corrosion of Materials by Lithium at Elevated Temperatures,” ORNL- 2674, Oak Ridge National Laboratory, March 1959 A.R. Keeton and C. Bagnall, Factors That Affect Corrosion in Sodium, Proceedings of the Second International Conference on Liquid Metal Technology in Energy Production, CONF-800401-P1, J.M. Dahlke, Ed., U.S. Department of Energy, 1980, p 7–18 to 7–25 B.H. Kolster, The Influence of Sodium Conditions on the Rate for Dissolution and Metal/ Oxygen Reaction of AISI 316 in Liquid Sodium, Proceedings of the Second International Conference on Liquid Metal Technology in Energy Production, CONF- 800401-P1, J.M. Dahlke, Ed., U.S. Department of Energy, 1980, p 7–53 to 7–61 J. Konys and H.U. Borgstedt, The Product of the Reaction of Alumina with Lithium Metal, J. Nucl. Mater., Vol 131, 1985, p 158–161 K. Natesan, Influence of Nonmetallic Elements on the Compatibility of Structural Materials with Liquid Alkali Metals, J. Nucl. Mater., Vol 115, 1983, p 251–262 D.L. Olson, P.A. Steinmeyer, D.K. Matlock, and G.R. Edwards, Corrosion Phenomena in Molten Lithium, Rev. Coatings Corros./Int. Q. Rev., Vol IV, 1981, p 349–434 S.J. Pawel, J.R. DeStefano, and E.T. Manneschmidt, Thermal Gradient Mass Transfer of Type 316L Stainless Steel and Alloy 718 in Flowing Mercury, J. Nucl. Mater., Vol 296, 2001, p 210–218 B.A. Pint, L.D. Chitwood, and J.R. DiStefano, Long-Term Stability of Ceramics in Liquid Lithium, J. Nucl. Mater., Vol 289, 2001, p 52–56 A.J. Romano, C.J. Klamut, and D.H. Gurinsky, “The Investigation of Container Materials for Bi and Pb Alloys, Part I: Thermal Convection Loops,” BNL-811, Brookhaven National Laboratory, July 1963 E. Ruedl, V. Coen, T. Sasaki, and H. Kolbe, Intergranular Lithium Penetration of Low-Ni, Cr-Mn Austenitic Stainless Steels, J. Nucl. Mater., Vol 110, 1982, p 28–36 J. Sannier and G. Santarini, Etude de la Corrosion de Deux Aciers Ferritiques par le Plomb Liquide Circulant dans un Thermosiphon; Recherche d'un Modele, J. Nucl. Mater., Vol 107, 1982, p 196–217 C.E. Sessions and J.H. DeVan, Thermal Convection Loop Tests of Nb-1%Zr Alloy in Lithium at 1200 and 1300 °C, Nucl. Appl. Technol., Vol 9, 1970, p 250–259 S.A Shields and C. Bagnall, Nitrogen Transfer in Austenitic Sodium Heat Transport Systems, Material Behavior and Physical Chemistry in Liquid Metal Systems, H.U. Borgstedt, Ed., Plenum Press, 1982, p 493–501 S.A. Shields, C. Bagnall, and S.L. Schrock, Carbon Equilibrium Relationships for Austenitic Stainless Steel in a Sodium Environment, Nucl. Technol., Vol 23, 1974, p 273–283 R.N. Singh, Compatibility of Ceramics with Liquid Na and Li, J. Am. Ceram. Soc., Vol 59, 1976, p 112–115

• •

• • •

•

D.L. Smith and K. Natesan, Influence of Nonmetallic Impurity Elements on the Compatibility of Liquid Lithium with Potential CTR Containment Materials, Nucl. Technol., Vol 22, 1974, p 392–404 A.W. Thorley, Corrosion and Mass Transfer Behaviour of Steel Materials in Liquid Sodium, Proceedings of the Third International Conference on Liquid Metal Engineering and Technology, Vol 3, The British Nuclear Energy Society, 1985, p 31–41 P.F. Tortorelli, Corrosion of Ferritic Steels by Molten Lithium: Influence of Competing Thermal Gradient Mass Transfer and Surface Product Reactions, J. Nucl. Mater., Vol 155–157, 1988, p 722–727 P.F. Tortorelli and J.H. DeVan, Effects of a Flowing Lithium Environment on the Surface Morphology and Composition of Austenitic Stainless Steel, Microstruct. Sci., Vol 12, 1985, p 213–226 P.F. Tortorelli and J.H. DeVan, Mass Transfer Kinetics in Lithium-Stainless Steel Systems, Proceedings of the Third International Conference on Liquid Metal Engineering and Technology, Vol 3, The British Nuclear Energy Society, 1985, p 81–88 J.R. Weeks and H.S. Isaacs, Corrosion and Deposition of Steels and Nickel-Base Alloys in Liquid Sodium, Adv. Corros. Sci. Technol., Vol 3, 1973, p 1–66

M. Ziomek-Moroz, Introduction to Corrosion for Constructive Purposes, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 135–138

Introduction Purposes

to

Corrosion

for

Constructive

Małgorzata Ziomek-Moroz, Albany Research Center, U.S. Department of Energy

Introduction PRINCIPLES OF METALLIC CORROSION play a fundamental role in developing industrial processes that employ corrosion for constructive purposes. These processes range from the spontaneous processes in galvanic cells, such as batteries and fuel cells, that convert energy for external applications, to electrolytic cell processes, such as electropolishing, electrochemical machining, and electrochemical refining, that consume energy from external sources. Each of these processes consists of an electrochemical cell with an anode, cathode, and conductive medium or electrolyte. Each of these processes involves electrochemical oxidation and reduction reactions. The purpose of this introduction is to show the similarities that exist between galvanic and electrolytic cells, while noting that the net or overall reaction for each type of cell leads us to label the electrodes and processes differently. For this purpose, a relatively simple electrochemical system has been chosen, and the changes in kinetics that occur with differentially small potential changes around the equilibrium electrode potentials of two reversible electrodes, Cu and Ag, are examined. In doing this, it should be clear that while differences in the kinetics of the spontaneous galvanic cell and the electrolytic cell can themselves be made differentially small, the industrial processes that are born out of these processes are quite different. The two reversible electrode systems chosen were a copper electrode immersed in a copper sulfate solution at unit activity of Cu2+ and a silver electrode immersed in a silver nitrate solution at unit activity of Ag+. The choice of these electrodes (or half-cells) avoids the complexity that arises with many electrodes of commercial interest where oxides or other stable corrosion products form on the surface. It also avoids the complexity that arises with reactions involving chemical species from different redox systems, such as zinc in an aqueous solution where the reduction reaction can be the evolution of hydrogen or the reduction of dissolved oxygen. The solutions are connected by a porous barrier and are at an ambient temperature of 25 °C (77 °F). The electrochemical reactions at each electrode are (for copper electrode):

(Eq 1) (for silver electrode): (Eq 2) where Ia and Ic are now just labels indicating that some charges are moving.

M. Ziomek-Moroz, Introduction to Corrosion for Constructive Purposes, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 135–138 Introduction to Corrosion for Constructive Purposes Małgorzata Ziomek-Moroz, Albany Research Center, U.S. Department of Energy

Reversible Cell Potential At equilibrium, the following conditions exist at each electrode: copper electrode Ia,Cu = Ic,Cu = Io,Cu silver electrode Ia,Ag = Ic,Ag = Io,Ag where I0 is the exchange current for the copper and silver electrochemical reactions, respectively. There is a characteristic electrostatic difference at these electrodes called the electrode potential difference that can be measured with respect to the standard hydrogen electrode (SHE).

At equilibrium there is a zero net reaction or current at each electrode. When the electrodes are connected through a high-impedance device such as an electrometer and the electrolytes are connected via a porous barrier to form the electrochemical cell:

the net current in the external circuit is zero; that is, the cell current Icell = 0. The copper electrode is more active or negative to the more noble or positive silver electrode (Fig. 1). The reversible cell potential (ecell,rev) is determined by the electrode potentials:

This potential could be measured with a high- impedance electrometer between the two electrodes (Fig. 2). The cell is reversible because the electrode potentials remain at their equilibrium potential when perturbed by incrementally small shifts in the cell reactions to produce reversibly small net reactions in one direction or the other. It should be noted that unit activity is not a necessary condition of this example. The reversible electrode potential, Erev, at any other activity of the metal ion, a, is given by the Nernst equation: Erev = Eo - (2.303RT/nF) log a

(Eq 3)

where R is gas constant; F is the Faraday constant, 96,500 C/gram-equivalent; T is the absolute temperature, K; and n is the number of electrons involved in the electrode reaction.

Fig. 1 Potential versus log current plot for reversible cell with copper and silver electrodes

Fig. 2 Reversible cell with copper and silver electrodes. Voltage is measured with a high-resistance electrometer (V), and current is measured with a zero resistance ammeter, (A).

M. Ziomek-Moroz, Introduction to Corrosion for Constructive Purposes, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 135–138 Introduction to Corrosion for Constructive Purposes Małgorzata Ziomek-Moroz, Albany Research Center, U.S. Department of Energy

Irreversible Cell Potential An electrode becomes irreversible when the electrode reactions are displaced from equilibrium and the electrode potential is no longer at the equilibrium potential. This happens in spontaneous processes when the cell is converting energy for external applications. It happens in electrolytic cells when energy from an external source is doing work on the electrodes. It happens in corrosion because the anodic and cathodic processes are coupled (or shorted) at a mixed or corrosion potential defining the kinetics of the corrosion reaction. Galvanic Cells. The example of the copper and silver electrodes functions as a galvanic cell when there is a net flow of electrons through an external circuit from the copper electrode to the silver electrode. When this occurs the following conditions exist at the electrodes: copper electrode Ia,Cu > Ic,Cu silver electrode Ia,Ag < Ic,Ag and the electrode potentials, E, shift to values between those of the initial equilbrium values and are no longer reversible:

The shift in potential, known as electrode polarization, is measured by the overpotential, η, which is the value by which the electrode potential has shifted from its equilibrium or reversible value (Fig. 3): η = E - Erev The effect of overpotential on the rate of the electrode reactions controlled by activation polarization is given by the Butler-Volmer equation. This corresponds to the straight lines in Fig. 1 and 4 that describe the anode and cathode reaction rates as a function of potential (or overpotential).

Fig. 3 Potential versus log current plot relating overpotential to potential scale for anodic and cathodic reactions. M is a metal, and n is a positive integer.

Fig. 4 Potential versus log current plot for galvanic cell with copper and silver electrodes The set of lines in Fig. 4 allows us to conveniently describe anodic and cathodic reaction rates at the copper and silver electrodes as a function of potential. The shift in electrode potentials for the galvanic cell gives a cell potential, ecell, that is less than the reversible cell potential of Fig. 1:

where the η-values are the overpotentials for the individual anodic and cathodic electrode reactions. Figure 5 is a circuit that allows the cell potential to be varied in a controlled manner from the reversible cell potential. By reducing the potential between direct current (dc) below ecell,rev, the spontaneous galvanic reaction is allowed to proceed. Since the anode reaction is larger than the cathode reaction at the copper electrode, the net reaction is: Cu → Cu2+ + 2eThe cathode reaction is larger than the anode reaction at the silver electrode, and the net reaction is: Ag+ + e- → Ag The overall cell reaction that spontaneously occurs then is: Cu + 2Ag+ → Cu2+ + 2Ag and the cell current, Icell ≠ 0. The galvanic cell can be described as a constructive galvanic cell when ecell > 0 and the call is delivering energy to an external device such as an electronic circuit, motor, or light.

Fig. 5 Galvanic cell with copper and silver electrodes is connected to an external power supply through a voltage divider. The potential, VC-D, can be varied. Voltage, V, is measured with a high resistance electrometer; current, A, is measured with a zero resistance ammeter. Following the 1953 International Union of Pure and Applied Chemistry (IUPAC) convention, since oxidation occurs at the copper electrode, it is the anode. Reduction occurs at the silver electrode, and it is the cathode. In terms of polarity, the copper electrode is the negative or more active electrode; the silver electrode is the positive or more noble electrode. The flow of current (positive charge) in the external circuit is from the silver electrode to the copper electrode; the electron flow is in the opposite direction, from the copper electrode to the silver electrode. Electrolytic Cells. The overall electrolytic cell process is not spontaneous because external work must be done on the system to produce the desired reaction. In the case of the copper and silver electrodes, this work is the deposition of copper instead of silver. There is a net flow of electrons through an external circuit from the silver electrode to the copper electrode. When this occurs, the following conditions exist at the electrodes: copper electrode Ia,Cu < Ic,Cu silver electrode Ia,Ag > Ic,Ag and the electrode potentials, E, shift to values outside the range of the initial equilibrium values and are no longer reversible:

This shift in electrode potentials for the electrolytic cell gives a cell potential, ecell, that is greater than the reversible cell potential:

as in Fig. 6 compared to Fig. 1. Since the cathode reaction is larger than the anode reaction at the copper electrode, the net reaction is: Cu2+ + 2e- → Cu The anode reaction is larger than the cathode reaction at the silver electrode, and the net reaction is: Ag → Ag+ + eThe overall cell reaction that occurs by the input of external energy is: Cu2+ + 2Ag → Cu + 2Ag+ and the cell current, Icell ≠ 0. By increasing the potential between points D and C in Fig. 5 above ecell,rev, the nonspontaneous electrolytic reaction proceeds in a controlled way, with the current now in the opposite direction from that of the galvanic cell and the cell potential greater than the reversible cell potential. Since oxidation occurs at the silver electrode, it is now the anode. Reduction occurs at the copper electrode, and it is the cathode. In terms of polarity, the copper electrode is still the negative or more active electrode, the silver electrode is still the positive or more noble electrode. The flow of current (positive charge) in the external circuit is from the copper electrode to the silver electrode; the electron flow is in the opposite direction, from the silver electrode to the copper electrode.

Fig. 6 Potential versus log current plot for electrolytic cell with copper and silver electrodes. The impressed potential difference is ecell. Corrosion Cells. The corrosion cell is a special case of a galvanic cell. The two electrodes are shorted together in a spontaneous process with both electrodes at the same potential, a mixed potential known as the corrosion potential (Ecorr). Corrosion cells differ from constructive galvanic cells in that no electrical energy is produced for external application. While there is a net flow of electrons through the circuit from the copper electrode to the silver electrode, this flow produces no useful external work. The following conditions exist at the electrodes: copper electrode Ia,Cu > Ic,Cu silver electrode Ia,Ag < Ic,Ag and the electrode potentials, E, shift to values between those of the initial equilbrium values. They are no longer reversible; that is;

In fact, they shift to the corrosion potential (Ecorr) where: The cell potential (Fig. 7) is:

Since the anode reaction is larger than the cathode reaction at the copper electrode, the net reaction is: Cu → Cu2+ + 2eThe cathode reaction is larger than the anode reaction at the silver electrode, and the net reaction is: Ag+ + e- → Ag The overall cell reaction that spontaneously occurs is: Cu + 2Ag+ → Cu2+ + 2Ag

Fig. 7 Potential versus log current plot for corrosion cell with copper and silver electrodes The cell current (Icell), when measured by a zero resistance ammeter, is equivalent to the corrosion current, Icorr. Since oxidation occurs at the copper electrode, it is the anode. Reduction occurs at the silver electrode, and it is the cathode. While both electrodes are at the same potential, the copper electrode is the more active electrode and the silver electrode is the more noble electrode. The flow of current (positive charge) in the external circuit is from the silver electrode to the copper electrode; the electron flow is in the opposite direction, from the copper electrode to the silver electrode. Table 1 summarizes the conditions that exist at the Cu and Ag electrodes in our example for the different types of cells: reversible, galvanic, electrolytic, and corrosion.

Table 1 Conditions at each electrode in example Cell type

Electrode reactions Cell potential Electrode potentials, ESHE Cu Ag Cu2+/Cu Ag+/Ag Ia = Ic = I0 Ia = Ic = I0 ecell,rev = 0.459 V Eo = 0.340 Eo = 0.799 Reversible Ia > Ic Ia < Ic ecell < ecell,rev E > Eo E < Eo Galvanic o Ia < Ic Ia > Ic ecell > ecell,rev E<E E > Eo Electrolytic Ia > Ic Ia < Ic ecell = 0 E > Eo E < Eo Corrosion Table 2 summarizes the conditions that exist for commercial and industrial electrochemical processes that employ corrosion for constructive purposes, as well as the conditions for corrosion. Table 2 Cell conditions for commercial and industrial electrode processes Cell type

Cell potential Reversible ecell,rev ecell < Galvanic ecell,rev Electrolytic ecell > ecell,rev Corrosion

ecell = 0

Net current I=0 I ≠ 0 (spontaneous)

Polarity Process Anode Cathode + … + Batteries, fuel cells

I ≠ 0 (impressed current or impressed potential)

+

-

Icorr (spontaneous)

-

+

Electrolytic polishing, electrochemical machining, electrochemical refining Corrosion, chemical-mechanical planarization

M. Ziomek-Moroz, Electropolishing, Corrosion: Fundamentals, Testing, and Protection,, Vol 13A, ASM Handbook, ASM International, 2003, p 139–142

Electropolishing M. Ziomek-Moroz, U.S. Department of Energy, Albany Research Center

Introduction ELECTROLYTIC POLISHING, also known as electropolishing, is an electrochemical process that involves anodic dissolution of a metal specimen (anode electrode) in an electrolytic cell. An electrochemical cell is a combination of an anode (an electronic conductor), an electrolyte (an ionic conductor), and a cathode (an electronic conductor); electrochemical processes occur with the passage of electric current in the cell (Ref 1). Irregularities on the anode surface are dissolved preferentially so that the surface becomes smooth and bright (Ref 2). Due to requirements for surface quality, electropolishing is a specific case of anodic dissolution and is essentially the reverse of electroplating. Electropolishing is widely used in many branches of science for purposes such as specimen preparation for corrosion studies. Commercial applications include improving appearance and reflectivity, improving corrosion resistance, removing edge burrs produced by mechanical cutting tools, removing the stressed and disturbed layer of metal caused by mechanical action (mechanical cutting), improving surfaces of medical tools, and removing radioactive contamination from surfaces (Ref 3, 4, 5).

References cited in this section 1. G. Wranglén, An Introduction to Corrosion and Protection, Chapman & Hall, 1985

2. “Metallography Principles and Procedures,” Leco Corporation 3. 45th Metal Finishing for Guide Book and Directory, Metals and Plastics Publications, Inc., 1977 4. K.P. Rajurkar, J. Kozak, and A. Chatterjee, Nonabrasive Finishing Methods Surface Engineering, Vol 5, ASM Handbook, ASM International, 1994, p 110–117 5. J.R. Davis, Surface Engineering of Carbon and Alloy Steels, Surface Engineering, Vol 5, ASM Handbook, ASM International, 1994, p 710

M. Ziomek-Moroz, Electropolishing, Corrosion: Fundamentals, Testing, and Protection,, Vol 13A, ASM Handbook, ASM International, 2003, p 139–142 Electropolishing M. Ziomek-Moroz, U.S. Department of Energy, Albany Research Center

Electrical Circuits Electropolishing can be carried out in two- electrode and three-electrode systems. A two- electrode laboratory setup is shown in Fig. 1. A workpiece or working electrode (WE), the anode, is connected to the positive terminal of a direct current (dc) power supply. The counterelectrode (CE), the cathode, is attached to the negative terminal. The variable resistor provides control of the current, which is monitored by the ammeter. In commercial use, a pulsed dc may be used.

Fig. 1 Two-electrode laboratory counterelectrode; dc, direct current

setup

for

electropolishing.

WE,

working

electrode;

CE,

In a three-electrode system (Fig. 2), a potentiostat regulates the dc power to the specimen (WE) and counterelectrode (CE) and receives information on the potential from the reference electrode (RE). The anode workpiece is connected to the working electrode terminal, and the cathode made of platinum or graphite is connected to the counterelectrode terminal. Finally a reference electrode, such as a saturated calomel electrode (SCE), is connected to the reference- electrode terminal to measure the potential of the workpiece.

Fig. 2 Three-electrode laboratory setup counterelectrode; RE, reference electrode

for

electropolishing.

WE,

working

electrode; CE,

The electrical potential of the power supply or potentiostat causes electronic conduction to the WE and CE and ionic conduction in the electrolyte. This may result in controlled anodic dissolution of the anode material and cathodic deposition on the CE of some species present in the electrolyte. During the electrolytic process, the products of the anodic metal dissolution react with the electrolyte to form a film at the surface of the metal as shown in Fig. 3.

Fig. 3 Formation of anodic film during electropolishing


Anodic Processes Anodic dissolution processes are complex. Depending on the nature of the dissolving metal (M), the electrolyte composition, and the current density, the following anodic (oxidation) reactions may occur: •

Transfer of metal ions into the electrolyte: Me → Me2+ + 2e-

•

Formation of oxide layers: Me + 2OH- → MeO + H2O + 2e-

•

(Eq 1)

(Eq 2)

Evolution of oxygen: 4OH- → O2 + 2H2O + 4e-

(Eq 3)

During electrolytic polishing, a voltage is applied between the anode and cathode, which are immersed in the electrolyte, through an external electrical circuit as shown in Fig. 1 and 2. If the current is monitored as a function of the cell voltage, polarization curves or current-voltage curves can be generated. Figure 4 shows an idealized anodic polarization curve with regions of cell voltage within which the reactions shown in Eq 1, 2, and 3 may occur.

Fig. 4 Idealized anodic polarization curve useful for electropolishing of materials showing active-passive behavior During electrolytic polishing, positive metal ions (cations) leave the specimen surface and diffuse into the electrolyte, and current increases with increasing applied voltage. In voltage region A-B, active dissolution or a direct etching takes place (the reaction represented by Eq 1). In the B-C voltage region, current decreases with increasing voltage, indicating unstable conditions on the metal surface. This represents active-passive behavior. Constant current with increasing voltage is observed in the C-D range when the anodic film is formed (the reaction shown in Eq 2). This represents passive behavior. At higher voltages (in the D-E range), current increases and the evolution of O2 takes place (the reaction shown in Eq 3), which resembles transpassive behavior. This type of curve, showing active, active-passive, passive, and transpassive regions, is obtained for copper electropolished in orthophosphoric acid. Electropolishing is achieved by operating at the mass-transferlimited rate of dissolution in the C-D region. The mass-transfer rate is represented by the limiting current density, iL, which depends on electrolyte concentration, viscosity, density, and stirring of the electrolyte, as well as the diffusivity of the ionic species present. Figure 5 is a polarization curve with two distinctive regions, namely: (a) film formation at low voltage and low current and (b) electrolytic polishing at higher voltage and current. Aluminum electropolished in perchloric acid exhibits this type of behavior. The difference in the shapes of the polarization curves for aluminum and copper indicate different electropolishing mechanisms.

Fig. 5 Idealized anodic polarization curve for electropolishing of materials in oxidizing electrolytes


Selection of Electrolytes Anodic dissolution of a metal depends on its chemical and physical properties, surface state, the composition of an electrolyte, and process conditions such as temperature, stirring, and current density (Ref 6). For materials that exhibit active-passive behavior, the most suitable polishing range is the passive region (C-D of Fig. 4). During etching, metal dissolves into the electrolyte while during anodic oxidation, the oxide film is not dissolved. The thickness of the anodic film or passivating layer may change, depending on current and electrolyte composition. The ideal electrolyte for electropolishing (Ref 7) must: • • • • • • • •

Provide high quality polishing at low voltages and current densities Have the ability to function over a large range of current densities and temperatures Offer stability and long service life Not dissolve the metal when no current is flowing; that is, no spontaneous corrosion occurs Be inexpensive, readily available, and safe Be recyclable Have an ohmic resistance (IR drop) that is sufficiently low to obtain the desired current density at low voltage Provide good throwing power; that is, a sample with complex geometry should dissolve uniformly over the entire surface

References cited in this section 6. P.V. Shigoev, Electrolytic and Chemical Polishing, Freund Publishing House, Ltd. 7. J. Flis, Corrosion of Metals and Hydrogen- Related Phenomena, Institute of Physical Chemistry, Polish Academy of Sciences, Warsaw, 1979


Kinetics of Electropolishing The kinetics of electropolishing is influenced by polishing current, temperature of electrolyte, polishing time, surface preparation, and agitation or stirring of electrolyte, which removes products and bubbles from the anode. The cathode area must be larger than the anode area in order to obtain an optimal polishing rate. Since the electropolishing process is a diffusion-controlled process, stirring or rotating the specimen significantly increases the rate of reaction. Under steady-state conditions, anodic reaction products accumulate on the surface and may interfere with the polishing process (Ref 4). Stirring or the use of a rotating disk electrode maintains a stable temperature and homogeneous composition of the electrolyte and prevents the buildup of corrosion products on the surface of the workpiece. The hydrodynamics of the electrolyte play a critical role in the diffusion control process characteristic of electropolishing as the transfer of reactants to the bulk solution is the rate-determining step in the electropolishing process (Ref 7). Mechanism of Electropolishing. Surface roughness during electropolishing can vary significantly depending on the stability of the film formed during the passivation stage. Cavities and crevices on the metal surface (Ref 6) require much higher current densities to obtain a uniform polished surface. During electropolishing, a corrosion-product layer forms on the metal surface that has a higher viscosity than the bulk fluid (Ref 8). The thickness of this layer varies over the anode surface area and is greater in crevices. Greater thickness is a result of higher current density. The leveling of the peaks and crevices of a rough surface follow a sequence where most peaks are dissolved in all directions while crevices are dissolved preferentially in one direction. The diffusion process during electropolishing is faster at the peaks than in the crevices. Unwanted selective etching, once a uniform surface is achieved, can be avoided by optimizing the electropolishing process conditions. Defects in Electropolishing. Pitting can occur during electropolishing as a result of incorrect electrolyte composition or due to the formation of oxygen bubbles that attack the surface being polished. Blemishes can occur due to intermetallic compounds or inherent defect on the metal surface. Scratches may remain on the electropolished face after inadequate surface preparation (Ref 6, 9). Hydrodynamic effects, kinematic viscosity, and concentration of electrolyte are the main variables that affect the removal rate during planarization. A rotating disk electrode is useful in the study of the hydrodynamic effects on an electrode. Rotational velocity is limited, so laminar flow is maintained. Generally, the rotating disk electrode is superior to static systems for study of the kinetics of heterogeneous reactions. For electrolytes in a diffusion controlled process where a stationary condition exists and the process is dependent on mass transfer of solute in a liquid, the following equation governs the process: V·

C = DΔC

(Eq 4)

where D is diffusivity; ΔC is concentration gradient; and V is liquid velocity. The diffusion layer is also controlled by removal of solute due to the electrochemical reaction and the boundary layer due to the changing velocity of the electrolyte. The following equations proposed by Levich describe the system in a flow pattern (laminar flow) using a rotating disk electrode (Ref 7): (Eq 5) and δ = 1.61D1/3ν1/6ω-1/2

(Eq 6)

where δ is diffusion layer thickness; J is the magnitude of diffusion flux; D is diffusion coefficient of a reactant, C0; C0 is concentration at the disk surface; C1 is concentration in bulk; ν is kinematics viscosity; and ω is angular velocity of disk. If the rate of reactant transport is slower than the rate of the chemical reaction, then C1 = 0 and the reaction occurs under diffusion controlled conditions. In the case where reaction rate is slower than the rate of transport, C1 = C0 and the process is under activation control where corrosion film is not present and the metal continuously corrodes. For a smooth disk in a laminar flow regime, the rotating disk electrode Reynolds numbers may be calculated by (Ref 7, 10) using the following equation: Re = r2ω/ν

(Eq 7)

where r is the radius of the disk. In a diffusion process, where C1 = 0, the rate of transfer of any reactant is greatly lower than the chemical change at interface. Combining Eq 5 and 6, one obtains: J = 0.62C0D2/3ν-1/6ω1/2

(Eq 8)

If one multiplies Eq 8 by a number of electrons participating in the reaction (n) and Faraday's constant (F), the removal rate current (I) in the electropolishing process can be obtained with: I = 0.62nFC0D2/3ν-1/6ω1/2

(Eq 9)

Physical properties of the electrolyte such as viscosity, density, concentration of solute and diffusivity play an important role in electrolytic polishing. The removal current density, (i = I/ area) is a linear relationship with the square root of angular speed of rotating electrode, ω.

References cited in this section 4. K.P. Rajurkar, J. Kozak, and A. Chatterjee, Nonabrasive Finishing Methods Surface Engineering, Vol 5, ASM Handbook, ASM International, 1994, p 110–117 6. P.V. Shigoev, Electrolytic and Chemical Polishing, Freund Publishing House, Ltd. 7. J. Flis, Corrosion of Metals and Hydrogen- Related Phenomena, Institute of Physical Chemistry, Polish Academy of Sciences, Warsaw, 1979 8. H.F. Walton, J. Electrochem. Soc., 1950, p 219–226 9. J.C. Scully, The Fundamentals of Corrosion, 3rd ed., The University of Leeds, United Kingdom, Pergamon Press, 1990, p 107–120 10. M.G. Foud., F.N. Zein. and M.L. Ismail, Electrochim. Acta, Vol 16, 1971, p 1477–1487


Properties of Electropolished Surfaces

The physical, corrosion, and mechanical properties of a polished metal surface depend on the nature and the surface state of the metal (Ref 6). The main variables are physical characteristics (crystalline lattice structure, stresses, strains), chemical nature, and surface geometry (distribution of peaks and crevices). Mechanically polished surfaces may have an amorphous or mechanically deformed layer that influences its corrosion behavior. However, this layer is gradually removed to the crystalline structure during electropolishing. Monitoring the chemical composition of the surface using scanning electron microscopy is a way to analyze surface heterogeneity and its effect on electropolishing. Effect of Electropolishing on Properties of Metals. In general, some of the mechanical properties of electropolished surfaces, such as hardness and Young's modulus, are increased. In some materials, such as stainless steel, no change is observed in static strength parameters (Ref 6). An effect on the fatigue behavior of metals is difficult to determine (Ref 6, 8) due to the complex nature of failure involving stress concentration, cracks, and surface defects. Mechanical polishing produces a cold-worked surface layer that generally enhances the fatigue life of metals. Fatigue life may be reduced when this layer is removed by electropolishing. On the other hand, surface roughness of the metal is reduced, which may increase the fatigue life of steel (Ref 6). Stress concentrations are also reduced. Magnetic properties of steel alloys are altered when the cold-worked layer is removed by electropolishing. Surface conductance is improved. The effect of electrolytic polishing on the corrosion properties of metals depends on the post- polished oxide that forms and how porous the protective film is. Resistance to uniform corrosion is generally improved following polishing (Ref 7). In the case of aluminum, the protective oxide film that is formed has proven to be beneficial.

References cited in this section 6. P.V. Shigoev, Electrolytic and Chemical Polishing, Freund Publishing House, Ltd. 7. J. Flis, Corrosion of Metals and Hydrogen- Related Phenomena, Institute of Physical Chemistry, Polish Academy of Sciences, Warsaw, 1979 8. H.F. Walton, J. Electrochem. Soc., 1950, p 219–226


References 1. G. Wranglén, An Introduction to Corrosion and Protection, Chapman & Hall, 1985 2. “Metallography Principles and Procedures,” Leco Corporation 3. 45th Metal Finishing for Guide Book and Directory, Metals and Plastics Publications, Inc., 1977 4. K.P. Rajurkar, J. Kozak, and A. Chatterjee, Nonabrasive Finishing Methods Surface Engineering, Vol 5, ASM Handbook, ASM International, 1994, p 110–117 5. J.R. Davis, Surface Engineering of Carbon and Alloy Steels, Surface Engineering, Vol 5, ASM Handbook, ASM International, 1994, p 710

6. P.V. Shigoev, Electrolytic and Chemical Polishing, Freund Publishing House, Ltd. 7. J. Flis, Corrosion of Metals and Hydrogen- Related Phenomena, Institute of Physical Chemistry, Polish Academy of Sciences, Warsaw, 1979 8. H.F. Walton, J. Electrochem. Soc., 1950, p 219–226 9. J.C. Scully, The Fundamentals of Corrosion, 3rd ed., The University of Leeds, United Kingdom, Pergamon Press, 1990, p 107–120 10. M.G. Foud., F.N. Zein. and M.L. Ismail, Electrochim. Acta, Vol 16, 1971, p 1477–1487



• • • •

R. Contolini, S. Mayer, M. Ziomek-Moroz, and E. Patton, Electropolishing Study for ULSI Planarization, Electrochemical Science and Technology of Copper, P. Vanysek, M. Alodan, J. Lipkowski, and O.M. Mangussen, Ed., Proc. vol PV 2000-30, Electrochemical Society, Pennington, 2002 J. Edwards, The Mechanism of Electropolishing of Copper in Phosphoric Acid Solution, J. Electrochem. Soc., 1953, p 223C O. Piotrowski, C. Madore, and D. Landolt, The Mechanism of Electropolishing of Titanium in Methanol-Sulfuric Acid Electrolytes, J. Electrochem. Soc., 1998, Vol 145, p 2302 Machining Methods, Electrochemical, Encyclopedia of Chemical Technology, p 608 H.F. Walton, The Anode Layer in the Electrolytic Polishing of Copper, J. Electrochem. Soc., 1950, p 219

V.K. Jain, Electrochemical Machining, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 143–152

Electrochemical Machining V.K. Jain, Indian Institute of Technology at Kanpur (India)

Introduction SALT SOLUTIONS (ELECTROLYTES)— unlike metallic conduction where only electrons are the charge carriers—conduct electrical charge by migration of ions in the medium. A feature of electrolysis is that electrical energy is used to produce an electrochemical reaction. The machining process based on this principle is known as electrochemical machining (ECM). In ECM, a small direct current electric potential (5 to 30 V) is applied across two electrodes, the cathode (tool) and anode (workpiece), immersed in an electrolyte. The

transfer of electrons between ions and electrodes completes the electrical circuit. Metal is detached, atom by atom, from the anode surface and appears in the electrolyte as positive ions, which react with hydroxyl ions to form metal hydroxides. During the process, H2 gas is evolved at the cathode (Fig. 1a). When the workpiece is steel, the reactions are: Fe + 2H2O → Fe(OH)2 + H2↑

(Eq 1)

2H2O + Fe(OH)2 + O2 → Fe(OH)3

(Eq 2)

Fig. 1 Elements of electrochemical machining (ECM). (a) Diagram showing dynamics. (b) Tool and work before ECM and after ECM. The density of parallel lines indicates current density. The smaller the interelectrode gap (IEG)— that is, the gap between anode and cathode—the greater the current flow (Fig. 1b) and the greater the metal-removal rate (MRR). The reaction products (metal hydroxides and gas bubbles) act as a barrier to the flow of electrolytic current. Their effect is minimized by supplying the electrolyte at a pressure varying in the range 2 to 35 kg/cm2 (30 to 500 psi), leading to the electrolyte flow velocity of 20 to 60 m/s (65 to 195 ft/s). The optimal pressure and the resulting electrolyte flow velocity will depend on the size of the tool and workpiece and the IEG. Electrolyte flowing at high velocity dilutes electrochemical reaction products and removes them from the IEG. It dissipates heat and limits the concentration of ions at the electrode surface to give higher machining rates. The electrolyte conductivity, the voltage across IEG, the gap itself, and tool shape are controlled to define the final anode (workpiece) profile. The electrolyte electrical conductivity depends on composition, concentration, and temperature. Electrolyte temperature can be controlled (within ±1° C) by heating or cooling the electrolyte in the tank. Feed to the tool should be provided at the same rate at which the workpiece surface is descending, so that the IEG remains almost constant during the process. This is known as machining under equilibrium conditions. Electrochemical machining principles have been used to perform a variety of machining operations (Ref 1), for example, turning, drilling, deburring, wire cutting, deep-hole drilling, and grinding. Electrochemical machining can be used to machine difficult-to-machine electrically conductive materials having complex shapes, deep holes with very high aspect ratio (up to 300:1), and three-dimensional profiles such as turbine blades. It is used in aerospace, automobile, and die-making industries.

Reference cited in this section 1. V.K. Jain, Advanced Machining Processes, Allied Publishers, Delhi, India, 2002

V.K. Jain, Electrochemical Machining, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 143–152 Electrochemical Machining V.K. Jain, Indian Institute of Technology at Kanpur (India)

Electrochemical Machining System An ECM system consists mainly of four subsystems: power source, electrolyte cleaning and supply system, tool and tool-feed system, and workpiece and workpiece-holding system (Fig. 2).

Fig. 2 Diagram of electrochemical machining systems The power supply provides direct current as high as 40,000 A at low electric potential. Silicon-controlled rectifiers (SCR) are used for control because of their rapid response to the load and their compactness. Power is cut off within 10 μs in the event of sparking to avoid any damage to the tool and workpiece. The electrolyte supply and cleaning system consists of a pump, filters, piping, control valves, heating or cooling coils, pressure gages, and one or more storage tanks. Electrolyte supply ports can be made in the tool, work, or fixture, depending on the requirements of the electrolyte flow. Electrolyte is cleaned by filters made of noncorroding materials such as Monel and stainless steel to avoid any blockage of the IEG, which is typically 0.3 to 1.0 mm (0.012 to 0.04 in.) by particles carried in the recycled electrolyte. Metallic piping should be grounded to prevent damage by corrosion. Tables and fixtures should also be grounded or should carry a cathodic potential to prevent corrosion. Further, ECM machines must be designed to withstand the hydraulic pressure of the electrolyte, which tends to separate tooling from the workpiece. Tools and fixtures are made of materials resistant to the very corrosive environment of the system. High thermal and electrical conductivity is needed for tool material. Easy machining of tool material is equally important because dimensional accuracy and surface finish of the tool directly affects the workpiece accuracy and surface finish. Even a small defect on the surface of the tool may leave marks, scratches, nicks, burrs, or lines on the machined surface due to the disturbed flow of electrolyte and altered current density at such irregularities. Aluminum, brass, bronze, copper, and carbon are used. Titanium is used as a tool material when machining with an acid electrolyte that anodizes the cathode. To remove the plated deposits from the tool, the current is reversed periodically. Table 1 gives relative properties of metals used for making tools in ECM. Due to high thermal conductivity, copper and brass will be damaged less than stainless steel and titanium by short circuits. Copper tungsten is highly resistant to damage from short circuits because of its high melting point. Table 1 Ranking of properties of commonly used tool materials

Property

Tool (electrode) material Copper Brass Stainless steel Titanium 1 4 53 48 Electrical resistivity 1.1 1 1.9 1.1 Stiffness 6.0 8.0 2.5 1.0 Machinability 7.5 1.0 2.6 Thermal conductivity 25 Source: Ref 2 There are three types of tools used in ECM— bare tool, coated tool, and bit-type of tool (Ref 3). Bare tools are usually not recommended, especially in cases of electrochemical drilling where straight-sided deep holes are needed. Areas on the tool where ECM action is not required should be insulated; however, due to the difficulties associated with achieving effective coatings, bit type of tools (Ref 3) are used (Fig. 3). As there is no wear and tear of the tool under normal ECM, theoretically tool life is infinite.

Fig. 3 Anode profile during EC drilling and EC bit drilling processes using (a) bare tool, (b) coated tool, (c) bare bit tool. 1, bare tool; 2, coated tool; 3, bit-type tool; 4, work; 5, stagnation zone; 6, front zone; 7, transition zone; 8, side zone; 9, tapered side for bit-type tool; 10, tool bit; 11, Perspex tool bit holder; 12, electric current conducting wire. Rtc, tool corner radius; ref, electrolyte flow hole radius Tool feed is controlled either by a stepper motor or a servosystem. Simple machines have manual control, while advanced machines have multiaxis computer control of feed and process parameters. Workpiece. Electrochemical machining can only be used on electrically conductive workpiece materials. Proper cleaning of the workpiece is essential after ECM to prevent residue from hardening on it. A clean water rinse for corrosion-resistant metals and alkaline cleaning of corrosive materials (steel, cast iron, etc.) is recommended. The workpiece is cleaned with a mild hydrochloric acid solution before a clean water rinse. During ECM, hydrogen is evolved at the tool (cathode); therefore, the probability of hydrogen embrittlement of the workpiece (anode) is minimal. There is no effect of ECM on ductility, yield strength, ultimate strength, and microhardness of the machined component. However, the fatigue strength of conventionally premachined components is reduced after ECM because of the removal of layers having compressive residual stresses. Little information is available about the effect of microstructure on the performance of ECM process. However, grain size and insoluble inclusions also affect surface roughness. Insoluble inclusions can stall the ECM process. Intergranular attack and other defects (pitting, selective etching) induced during ECM seem to be secondary reasons for the reduction in fatigue strength. Intergranular attack can be controlled by appropriate selection of electrolyte and ECM parameters. No effect of hardness of the workpiece material on the process performance has been reported. Workpiece-holding devices are made of electrically nonconducting materials—such as glass fiber reinforced plastics, plastics, and Perspex— having good thermal stability and low moisture absorption properties.

References cited in this section 2. T.L. Lievestro, Electrochemical Machining, Machining, Vol 16, Metals Handbook, 9th ed., ASM International, 1989, p 533–541 3. V.K. Jain and P.C. Pandey, Investigations into the Use of Bits as a Cathode in ECM, Int. J. Mach. Tools Des. Res., Vol 22 (No. 4), 1982, p 341–352


Theory of ECM Electrochemical machining differs from other industrial processes based on Faraday's laws of electrolysis, such as electroplating. In ECM, the shape and size of the workpiece is changed in a controlled manner, and the minimum current density employed is higher than 8 A/mm2. Faraday's law of electrolysis is: (Eq 3)

where m is mass (grams) removed in time t, I is current, t is time (seconds), EW is gram equivalent weight of workpiece metal (EW = Aw/z; Aw = atomic weight, z = valency of dissolution), and F is Faraday's constant (96,500 coulombs/gram equivalent, or ampere-seconds/gram equivalent). Equation 3 does not account for the effects of some of the significant process variables, that is, overpotential, presence of passive film, variation in the electrical conductivity of the electrolyte due to temperature and void concentration along the electrolyte flow path and valence changes during electrochemical dissolution. Regarding the latter, dissolution of iron in NaCl solution depends on the machining conditions and may be either in the form of ferrous hydroxide (Fe2+) or ferric hydroxide (Fe3+) ions. The mode of dissolution during alloy machining is still more difficult to identify (Ref 1). Since the workpiece is usually an alloy rather than a single element, the evaluation of m becomes difficult because electrochemical equivalent of an alloy (EWa) is generally not well defined. However, two methods have been proposed (Ref 4) to solve this problem. Method 1 is the percentage by weight method. The value of EWa is calculated by multiplying the chemical equivalent of individual element (EWi) by their respective proportion by weight percent (Xi), and then summing them up as given by: (Eq 4) where n is number of the constituent elements in the alloy. Method 2 is the superposition of charge method. The amount of electrical charge required to dissolve the mass contribution of individual element of the alloy is calculated and then EWa is evaluated as: (Eq 5) These two methods do not give exactly the same value of EWa. The total current (I) is not used in dissolving material from the workpiece. Actual MRR depends on the current efficiency (η), which ranges from 75 to 100%.

References cited in this section 1. V.K. Jain, Advanced Machining Processes, Allied Publishers, Delhi, India, 2002 4. J.A. McGeough, Principles of Electrochemical Machining, Chapman and Hall, London, 1974


Metal-Removal Rate It is desirable to have minimum possible IEG (≤0.5 mm) for accurate reproduction of tool shape on the workpiece. The analysis (Ref 1, 5) given in this section is based on the following assumptions: • • •

Electrical conductivity of electrolyte during the process remains constant. The surfaces of anode and cathode are considered to be equipotential. Effective voltage (V - ΔV) across the electrodes remains constant.

•

The anode dissolves at one fixed valency of dissolution.

In the above assumptions, V is constant potential difference applied across the electrodes, and ΔV is sum of electrode overpotentials including reversible potential. The difference between the equilibrium and working values of the potentials is known as overpotential. This overpotential includes activation, concentration, and resistance overpotentials (Ref 4). In the case of ECM with plane parallel electrodes, MRR (in gram MRRg, in volume MRRv) can be calculated as: (Eq 6)

(Eq 7) where ρa is density of anode (workpiece). Table 2 gives theoretical material-removal rate for different metals (Ref 4), assuming current efficiency as 100%. The MRR depends on the valence of dissolution and atomic weight of the material. Theoretical MRR ranges from 11 to 84 mm3/s (2.4 to 18.5 in.3/h). Table 2 Theoretical metal-removal rate (MRR) for current of 1000 A

Aluminum Beryllium Chromium

Atomic (Aw) g 26.97 9.0 51.99

lb 0.595 0.198 0.115

g/cm3 2.67 1.85 7.19

Cobalt

58.93

0.130

8.85

Copper

63.57

0.140

8.96

Iron

55.85

0.123

7.86

Magnesium Manganese

24.31 54.94

0.054 0.121

1.74 7.43

Molybdenum 95.94

0.212

10.22

Nickel

58.71

0.129

8.90

Niobium

92.91

0.205

9.57

Silicon Tin

28.09 118.69

0.062 0.262

2.33 7.30

Titanium

47.9

0.106

4.51

Metal

weight Density (ρ)

Valence, z

lb/in.3 0.0965 3 0.0668 2 0.2598 2 3 6 0.3197 2 3 0.3237 1 2 0.2840 2 3 0.0629 2 0.2684 2 4 6 7 0.3692 3 4 6 0.3215 2 3 0.3457 3 4 5 0.0842 4 0.2637 2 4 0.1629 3 4

Metal-removal rate (MRRg) g/s lb/h 0.093 0.74 0.047 0.37 0.269 2.14 0.180 1.43 0.090 0.71 0.306 2.43 0.204 1.62 0.660 5.24 0.329 2.61 0.289 2.29 0.193 1.53 0.126 1.00 0.285 2.26 0.142 1.13 0.095 0.75 0.081 0.64 0.331 2.63 0.248 1.97 0.166 1.32 0.304 2.41 0.203 1.61 0.321 0.04 0.241 1.91 0.193 1.53 0.073 0.58 0.615 4.88 0.307 2.44 0.165 1.31 0.124 0.98

Metal-removal rate (MRRv) mm3/s in.3/h 35 7.7 25 5.5 38 8.4 25 5.5 13 2.9 35 7.7 23 5.1 74 16.3 37 8.1 37 8.1 25 5.5 72 15.8 38 8.4 19 4.2 13 2.9 11 2.4 32 7.0 24 5.3 16 3.5 34 7.5 23 5.1 34 7.5 25 5.5 20 4.4 31 6.8 84 18.5 42 9.2 37 8.1 28 6.2

Tungsten

183.85

0.405

19.3

Uranium

238.03

0.525

19.1

0.6973 6 8 0.6900 4 6 0.2576 2

65.37 0.114 7.13 Zinc Source: Ref 4 Linear metal-removal rate (or penetration rate, MRRl) is given by:

0.317 0.238 0.618 0.412 0.339

2.52 1.89 4.90 3.27 2.69

16 12 32 22 48

3.5 2.6 7.0 4.8 10.6

(Eq 8) where J is current density. The current density in the IEG is a function of shapes of the electrodes, the distance separating the electrodes (y), and effective voltage between them: (Eq 9) where k is electrical conductivity of electrolyte, y is IEG, and A is workpiece area under the tool. The voltage applied across the IEG is a controllable parameter and greatly influences ECM performance. For a constant voltage power supply, the magnitude of current flow is controlled by the dynamics of IEG, y.

References cited in this section 1. V.K. Jain, Advanced Machining Processes, Allied Publishers, Delhi, India, 2002 4. J.A. McGeough, Principles of Electrochemical Machining, Chapman and Hall, London, 1974 5. H. Tipton, The Dynamics of Electrochemical Machining, Advances in Machine Tool Design and Research, Pergamon Press, Birmingham, 1964, p 509–522


Interelectrode Gap During ECM, the tool moves toward the workpiece at the feed rate (f), but at the same time, work surface moves away from the tool at a rate equal to the penetration rate (MRRl). Therefore, the effective rate of change of IEG is given by: (Eq 10) Under equilibrium conditions (dy/dt = 0, IEG = ye), the equilibrium gap distance (ye) and the feed rate have the relationship: (Eq 11)

Knowing the feed rate, the equilibrium gap (ye) can be calculated from Eq 11. Equation 10 can be rewritten (assuming efficiency as 100%) as: (Eq 12) where c = (EW)(V - ΔV)k/Fρa; c is assumed to remain constant for a given workpiece and electrolyte combination only if (V - ΔV) and k remain constant during the process. In practice, two cases arise for which Eq 12 can be solved, that is, zero feed rate (f = 0) and finite feed rate (f ≠ 0). In case of zero feed rate, the solution (Ref 5) of Eq 12 is given by Eq 13 for the initial condition, y = y0 at t = 0. (Eq 13) For finite feed rate, the IEG attempts to attain the equilibrium gap. The solution (Ref 4, 5) of Eq 12 is given by: (Eq 14) where yt is IEG at any time t. Generalized equation for IEG. Equation 13 predicts an infinite gap at t = ∞. Equation 14 is an implicit equation that is solved for yt only by iterative process. Hence, Eq 14 is not especially convenient to handle in those cases where numerical methods such as finite-element methods, finite-difference methods, or boundary-element methods are applied to analyze the ECM process. A single equation has been derived (Ref 6) for numerical methods analysis. It is applicable simultaneously to both the cases, that is, for f = 0 and f ≠ 0. Equation 10 can be written in the following form: (Eq 15) All the factors except J on the right-hand side of this equation can be treated as constant during ECM. The domain of interest in numerical methods is then divided into elements whose size is quite small; therefore, within an element J can be assumed to remain constant over a small interval of time dt. Therefore, Eq 15 can be written as: dy = (c - f)dt

(Eq 16)

where c = {[(EW)η]/Fρa} and is assumed to be constant. After integration, Eq 16 can be written (for y = y0 at t = 0) as: y = y0 + (c - f)t

(Eq 17)

Equation 17 is valid only for a small interval of time (t = Δt). Hence, it can be written: y = y0 + (c - f)Δt

(Eq 18)

Equation 18 is valid for both the cases for feed rate (Ref 7), and the total computer time is reduced.

References cited in this section 4. J.A. McGeough, Principles of Electrochemical Machining, Chapman and Hall, London, 1974 5. H. Tipton, The Dynamics of Electrochemical Machining, Advances in Machine Tool Design and Research, Pergamon Press, Birmingham, 1964, p 509–522 6. V.K. Jain and P.C. Pandey, Tooling Design for ECM: A Finite Element Approach, J. Eng. Ind. (Trans. ASME), Vol 103, 1981, p 183–191 7. V.K. Jain and P.C. Pandey, Tooling Design for ECM, Prec. Eng., Vol 2, 1980, p 195–206


Electrolyte Conductivity Electrolyte electrical conductivity is a function of local temperature (T) and presence of sludge and gas bubbles (hydrogen, oxygen, vapor, and other gas bubbles) in the IEG. As a result, the IEG gets tapered along the electrolyte flow direction. For precision machining, temperature of the electrolyte should be controlled within ±1 °C (±1.8 °F). This may require heating or cooling the electrolyte. In the analysis given below, only the effects of temperature and hydrogen gas bubbles are taken into account. Using the law of conservation of heat, the temperature gradient (dT/dx) along the path of electrolyte flow (xdirection) can be derived (Ref 8) as follows: (Eq 19) where the density (ρe) and specific heat (Ce) of the electrolyte are assumed to remain constant in the IEG. Average values of IEG ( ) and electrolyte flow velocity (Ū) are also assumed to remain unchanged. Thus,

can be treated as constant. Electrical conductivity (k) of the electrolyte varies linearly with temperature as: k = k0(1 + α · ΔT)

(Eq 20)

where α is a temperature coefficient of specific conductance [α = dk/(k0dT)]. Solving after substituting Eq 20 in Eq 19 gives: (Eq 21) It can also be shown that: (Eq 22) To incorporate the effect of temperature as well as hydrogen gas bubbles on electrolyte conductivity, the following equation has been suggested (Ref 9, 10): k = k0{(1 + α · ΔT)(1 - αv)n}

(Eq 23)

A value of exponent n between 1.5 to 2.0 has been used (Ref 9, 10, 11). This equation, however, does not account for the effect of size and distribution of H2 bubbles. Effect of change in temperature and concentration of NaCl and HCl on their conductivity is shown in Fig. 4.

Fig. 4 Electrical conductivity of electrolyte as a function of temperature and concentration of (a) NaCl, (b) HCl. Source: Ref 12

References cited in this section 8. P.C. Pandey and H.S. Shan, Modern Machining Processes, Tata McGraw Hill, Delhi, India, 1980 9. J.F. Thorpe and R.D. Zerkle, Analytical Determination of the Equilibrium Gap in Electrochemical Machining, Int. J. Mach. Tools Des. Res., Vol 9, 1969, p 131–144 10. J. Hopenfeld and R.R. Cole, Prediction of the Dimensional Equilibrium Cutting Gap in Electrochemical Machining, J. Eng. Ind. (Trans. ASME), Vol 91, 1969, p 75–765 11. V.K. Jain, P.G. Yogindra, and S. Murugan, Prediction of Anode Profile in ECBD and ECD Operations, Int. J. Mach. Tools Manuf., Vol 27 (No. 1), 1987, p 113–134 12. A.E. DeBarr and D.A. Oliver, Ed., Electrochemical Machining, Macdonald and Co. Ltd, 1975


Process Control Electrochemical machining has a unique self- regulating feature. Under the equilibrium condition (f = MRRl), machining takes place at a constant IEG (that is, equal to ye). If this were not true at all times during ECM, either f < MRRl or f > MRRl.

If f < MRRl, initially IEG will increase and it will attain a value greater than the equilibrium gap value (y > ye). Because of this, the current density will also decrease as compared to Je (current density under equilibrium condition) and the MRRl will also decrease. In other words, the difference between f and MRRl will decrease, or the process will move toward the equilibrium gap condition. If f > MRRl, the reverse argument will be true, and again the IEG will converge to the equilibrium gap (or MRRl = f ). The self-regulating convergence is shown in the Fig. 5, where the IEG always attempts to reach the equilibrium gap value, ye.

Fig. 5 The interelectrode gap (IEG) as a function of time the self-regulating feature of ECM is shown, as gaps with different initial values (y) converge to the equilibrium gap (ye).


Electrolytes A good electrolyte for ECM must have high electrical conductivity, high specific heat, good chemical and electrochemical stability, controllable passivating effect, low viscosity, and low toxicity and corrosivity (Ref 2, 12). Electrolyte in the IEG conducts current between anode and cathode, removes reaction products, and dissipates heat produced during the process. The electrolytes perform better when the crystal structure of the workpiece is fine grained, and its constituents (in case of an alloy) dissolve at uniform rate. Electrolytes used in ECM are classified in three categories: neutral salt, alkaline, and acidic. Each type has its own merits. Sludging electrolytes produce hydroxide sludges that must be removed before recycling. Sodium chloride (NaCl) solutions are the most commonly used electrolytes in ECM. During ECM, water is depleted, producing H2 gas and hydroxides (Eq 1 and 2), which are insoluble in water. NaCl electrolyte is corrosive in nature and produces large amounts of sludge. The ratio of the amount of sludge formed to the amount of metal dissolved ranges between 6 and 8 for NaCl electrolyte. There is no universal electrolyte that can be employed for every type of metal and alloy. Table 3 (Ref 12) recommends some electrolytes for a limited range of metals and alloys. Under certain circumstances, mixtures of electrolytes, such as NaNO3 and NaCl or NaCl and HCl, are used. For instance, NaNO3 is passivating and more expensive, but gives smoother surfaces in certain cases.

Table 3 Electrolytes for the electrochemical machining of metals Workpiece metal

Electrolyte Major constituent

Removal rate Concentration (max) kg/L of lb/gal of H2O H2O 0.30 2.5 0.60 5 0.78 6.5 0.30 2.5 0.60 5 0.60 5 0.60 5 0.30 2.5 0.12 1 (g) 0.18 1.5(g) 0.18 1.5 0.30 2.5 0.30 2.5 0.60 5 0.30 2.5

mm3 × 103/min per 1000 A 2.1 2.1 2.0 2.0(a)(b) 2.0(a)(b) 1.6(c) 2.1 2.1 1.6 1.0 1.0 1.0 4.4 3.3 2.1

in.3/min per 1000 A 0.13 0.13 0.12 0.12(a)(b) 0.12(a)(b) 0.10(c) 0.13 0.13 0.10 0.06 0.06 0.06 0.27 0.20 0.13

NaCl or KCl NaNO3 NaClO3 NaCl NaNO3 NaNO3 White cast iron NaNO3 Aluminum and aluminum (d) alloys NaCl or KCl NaCl or KCl(e) Titanium alloys NaOH(f) Tungsten NaOH(h) Molybdenum NaCl or KCl (d) NaCl or KCl Copper and copper alloys NaNO3 NaCl or KCl Zirconium (a) Feed rates limited by graphite particle size. (b) Maximum; can vary widely. (c) Rough surface finish. (d) NaNO3 electrolyte provides better surface finish. (e) Voltage must be greater than. (f) NaOH used up in process and must be replenished. (g) Minimum of 9 kg/L (0.75 lb/gal.) (h) pH of electrolyte decreases with use; maintain pH by adding NaOH or KOH. Source: Ref 2 Nonsludging electrolyte such as NaOH solutions are used in ECM of heavy metals such as tungsten and molybdenum and their alloys. The electrolyte becomes depleted during ECM due to the formation of compounds such as sodium tungstate (Na2WO4 · 2H2O) and sodium molybdate (Na2MoO4 · 2H2O) while machining tungsten and molybdenum, respectively. These compounds are soluble in water (hence, no sludge), but these heavy metals tend to plate out onto the cathode, which is undesirable. However, a periodic voltage polarity reversal will keep the tool cathode surfaces clean. Filtration. Effective filtration of sludge-forming electrolyte is very essential to maintain the performance of the ECM process. In no case should the electrolyte contain more than 2 wt% sludge when in use. Presence of sludge particles in the IEG can lead to short circuits. Filtering, centrifuging, and settling are used to remove sludge. Electrolyte flow rate, controlled by pressure, in the IEG affects electrolyte conductivity by changing its temperature and the hydrogen gas bubble size and distribution in the IEG. Electrolyte flow rate must be high enough to remove machining by-products as quickly as possible. Electrolyte flow rate influences surface finish Ra value (centerline average value of surface finish is abbreviated as Ra) and current efficiency, η, as well as MRR. A high ratio of electrolyte flow rate to the current across the IEG is desirable, but the cost of pumping increases. A rule of thumb is approximately 1 L/min for each 100 A for machining steel with NaCl. High flow rate usually reduces the value of Ra and minimizes the formation of deposits on the cathode. Excessive flow rate can lead to local erosion of tool/ workpiece and sticking of precipitates on to the workpiece. To produce a smooth, uniform surface on the workpiece, the tool must be designed so there is uniform flow over the entire machining area. This is difficult at high flow velocities of 30 to 60 m/s (100 to 200 ft/s). The design of tools for complex-shaped workpieces requires an understanding of multiphase fluid flow (involving liquid, gas, and solid phases), electrical and electrochemical principles, as well as experience and ingenuity. Back Pressure. Attempts have been made (Ref 2) to study the effect of back pressure in the IEG. Increased back pressure in ECM reduces imprints of flow lines on the workpiece and increase drilled-hole diameter, but results Steel; iron- nickel-, and cobalt-base alloys Steel; hardened tool steel Gray iron

in higher hydraulic forces and tooling cost. Hole diameter and current at constant voltage are increased with increase in back pressure because the resulting hydrogen bubbles are smaller, leading to increased conductivity (Fig. 6).

Fig. 6 Effect of electrolyte back pressure on hole size and current at constant voltage. Source: Ref 2 Safety precautions are important in ECM. Dry oxidizing salts such as sodium nitrate and sodium chlorate are hazardous. Mists, vapors, and dusts of alkaline electrolyte can damage body tissues. Appropriate ventilation of the work area and use of protective gloves, face shields, and masks are required. To prevent hydrogen explosion, hydrogen evolved during ECM must be vented from the work area. Proper storage, handling, and disposal of the electrolytes used in machining of heavy metals are also very important.

References cited in this section 2. T.L. Lievestro, Electrochemical Machining, Machining, Vol 16, Metals Handbook, 9th ed., ASM International, 1989, p 533–541 12. A.E. DeBarr and D.A. Oliver, Ed., Electrochemical Machining, Macdonald and Co. Ltd, 1975 V.K. Jain, Electrochemical Machining, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 143–152 Electrochemical Machining V.K. Jain, Indian Institute of Technology at Kanpur (India)

Workpiece Shape Prediction Analytical and computational models (Fig. 7) for workpiece shape prediction have been developed (Ref 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28), ranging from the simple “cos θ” principle to those based on approximate numerical techniques. These models have been used to predict work profile in different zones of electrochemically drilled blind holes, that is, side, transition, front, and stagnation zones (Fig. 8). The ability to predict variations in IEG in different zones for any given tool and operating conditions is a prerequisite for proper design of ECM tools and anode shape prediction (Ref 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28).

Fig. 7 Various proposed models for anode shape prediction and tool design in ECM

Fig. 8 Diagram of electrochemical drilling with outward mode of electrolyte flow showing various zones. as, side gap Techniques for workpiece shape prediction require the use of high-speed computers. In cos θ method (Ref 5), equilibrium shape is computed corresponding to a tool whose profile is approximated by a large number of planar sections inclined at different angles. If the feed direction is inclined at an angle θ normal to the tool face, the equilibrium gap is equal to (ye/cos θ). This approach is based on many simplified assumptions and is not recommended for complex- shaped workpieces (Ref 6). Empirical methods attempt to derive equations based on the experimental observations of IEG in the transition and side zones to predict a complete anode profile. Various models (Ref 15, 16, 29) have been proposed in terms of equilibrium gap, tool corner radius (rc), and bare length of the tool (bb). Nomographic methods have been proposed (Ref 17, 29) for the evaluation of side gap for the known values of ye, rc, and bb. Such methods have proved useful in planning an ECM operation. However, the empirical and nomographic methods cannot be generalized and are valid only for the specified machining conditions, tool and work material combination, and type of electrolyte.

More complicated problems may require Laplace's equation to be solved to determine the operating potential, V, by numerical methods such as finite-difference method (Ref 14), finite- element method (Ref 11), or boundary-element method (Ref 19). The procedure followed to solve such problems is: • • • •

The electric field potential at different points is calculated by solving Laplace equation using one of the numerical methods. Using the field-distribution data, current density is calculated. With the help of Eq 21, temperature distribution in the IEG is evaluated, and it is used to modify the electrolyte conductivity. The interelectrode gap is computed using Eq 18 at different points in the domain of interest to predict the anode shape.

References cited in this section 5. H. Tipton, The Dynamics of Electrochemical Machining, Advances in Machine Tool Design and Research, Pergamon Press, Birmingham, 1964, p 509–522 6. V.K. Jain and P.C. Pandey, Tooling Design for ECM: A Finite Element Approach, J. Eng. Ind. (Trans. ASME), Vol 103, 1981, p 183–191 11. V.K. Jain, P.G. Yogindra, and S. Murugan, Prediction of Anode Profile in ECBD and ECD Operations, Int. J. Mach. Tools Manuf., Vol 27 (No. 1), 1987, p 113–134 13. H. Tipton, The Determination of Tool Shape for Electrochemical Machining, Mach. Prod. Eng., 1968, p 325–328 14. Report No. 145, Production Engineering Research Association, Melton Mowbray, 1968 15. V.K. Jain, Vinod K. Jain, and P.C. Pandey, Corner Reproduction Accuracy in Electrochemical Drilling of Blind Holes, J. Eng. Ind. (Trans. ASME), Vol 106, 1984, p 55–61 16. W. König and H.J. Hümb, Mathematical Model for the Calculation of the Contour of the Anode in Electrochemical Machining, Ann. CIRP, Vol 25, 1977, p 83–87 17. H. Heitmann, Surface Roughness Formation and Determination of Optimum Working Conditions in Electrochemical Die Sinking, Proc. Second AIMTDR Conf., 1968, p 41–47 18. Y.G. Tsuei, C.H. Yen, and R.H. Nilson, Theoretical and Experimental Study of Workpiece Geometry in Electrochemical Machining, ASME, Paper 76-WA/Prod, presented in WAM, 1977, p 1–5 19. O.H. Narayanan, S. Hinduja, and C.F. Noble, The Prediction of Workpiece Shape During Electrochemical Machining by the Boundary Element Method, Int. J. Mach. Tools Des. Res., Vol 26 (No. 3), 1986, p 323–338 20. R.H. Nilson and Y.G. Tsuei, Free Boundary Problem for the Laplace Equation with Application to ECM Tool Design, J. Eng. Ind. (Trans. ASME), Vol 98, 1976, p 54–58 21. P. Lawrence, Computer Aided Design for ECM Electrodes, Int. J. Mach. Tools Des. Res., Vol 21, 1981, p 379–385 22. M.S. Reddy, V.K. Jain, and G.K. Lal, Tool Design for ECM: Correction Factor Method, J. Eng. Ind. (Trans. ASME), Vol 110, 1988, p 111–118

23. V.K. Jain, K. Ravi, G.K. Lal, and K.P. Rajurkar, Investigations into Tool Design for Electrochemical Drilling, Process. Adv. Mater., Vol 1 (No. 1), 1991, p 105–121 24. O.H. Narayanan, S. Hinduja, and C.F. Noble, Design of Tools for Electrochemical Machining by the Boundary Element Method, Proc. Inst. Mech. Eng., Vol 200 (C3), 1986, p 195–205 25. V.K. Jain and K.P. Rajurkar, An Integrated Approach for Tool Design in ECM, Prec. Eng., Vol 13 (No. 2), 1991, p 111–124 26. M.B. Nanyakkara and C.N. Larsson, Computation and Verification of Workpiece Shape in Electrochemical Machining, 20th Int. Machine Tool Design and Research Conf., UMIST, U.K., 1976, p 617–624 27. H. Tipton, Calculation of Tool Shape for ECM, Electrochemical Machining, C.L. Faust, Ed., Electrochem. Soc., 1971, p 87–102 28. V.K. Jain and P.C. Pandey, On the Reproduction of Corner Anode Profile During Electrochemical Drilling of Blind Holes, J. Eng. Ind. (Trans. ASME), Vol 106 (No. 1), 1984, p 55–61 29. W. König and D. Pahl, Accuracy and Optimal Working Conditions in ECM, Ann. CIRP, Vol 8 (No. 2), 1970, p 223–230 V.K. Jain, Electrochemical Machining, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 143–152 Electrochemical Machining V.K. Jain, Indian Institute of Technology at Kanpur (India)

Tool Design Electrochemical machining tools used to be produced by experienced and skilled toolmakers using mostly hand finishing. Nowadays, computer-controlled machine tools can produce ECM tools more accurately while reducing time and cost. Tool design involves the computation of tool shape and size, which would produce a workpiece having the desired shape, size, and accuracy when machined under the specified machining conditions (Ref 2, 12, 13, 14). Tool-design problems are still difficult to solve. Cumbersome and expensive, “trial-and-error” philosophy is often exercised on the shop floor. Several attempts have been made to develop a useful tool- design model (Fig. 7) (Ref 13, 20, 21, 22, 23, 24, 25, 27). The cos θ (Ref 13) method has been developed for designing cathode shape, but it has many limitations. The complex variable technique (Ref 20) is applicable only for simple shapes and is unable to tackle discontinuities of work and tool surfaces. The boundary-element method (Ref 24) has also been employed for the solution of tool-design problems. A correction-factor method using finite-element analysis has been proposed (Ref 22, 23) for one- and twodimensional tool-design problems. The concept of correction factor has been applied to modify the tool shape in each cycle of computation. The tool-shape-modification process continues until the difference between the work shape obtainable by the designed tool and the desired work shape is well within the specified tolerances. The tool-design procedure involves five major steps: 1. Using the anode shape prediction model (based on finite-element analysis), the anode profile is estimated for an assumed (or modified) tool shape and specified machining conditions. The initial tool shape is assumed to be complementary to the workpiece shape.

2. The computed and the desired work shapes are compared, and the difference between the two is evaluated as an error. 3. If the error is more than the specified tolerance value, a correction factor is calculated. 4. The tool shape is modified by applying the correction factor. 5. Steps 1 to 4 are repeated until the desired tool shape is achieved. This concept of correction factor can be applied to other tool-design models employing finite-difference or boundary-element technique. Figure 9 shows the computed work shape and the work shape obtained experimentally for a bit- type tool. The difference between the experimental and the computed work shapes at the top of the drilled hole is due to the effect of stray- current attack and was not incorporated in the present model. The difference between the designed tool shape and the used tool shape is attributed to inaccuracy in the anode shape prediction model and the measurement of overcut in the transition zone. Obtainable anode shapes indicate that to achieve taper-free cavities, either insulated or a bit-type tool should be used. Overcut and taper are governed by the bare part of the tool (bb).

Fig. 9 Comparison of experimental and designed tool profiles used during electrochemical bit drilling, feed rate, f = 0.037 mm/min, electrolyte conductivity, k = 0.0007 Ω-1/mm. (a) Ev = 6.54 V, rt = 6.03 mm, rtc = 2.39 mm, bb = 5.22 mm. (b) Ev = 7.64 V, rt = 5.50 mm, rtc = 1.50 mm, bb = 3.80 mm. Source: Ref 22 Different modes of electrolyte flow are shown in Fig. 10. Sudden changes in flow direction may lead to stagnation resulting in striations, ridges, or protuberances on the workpiece. This predicament can be resolved by having more than one hole (or slot) in the tool to deliver electrolyte to the IEG, but sometimes it is necessary to conduct an additional ECM operation to get rid of these irregularities.

Fig. 10 Tools for electrochemical machining. (a) Dual external-cutting tool for a turbine blade, crossflow type. Special fixtures are to confine electrolyte flow. (b) Tool for sinking a stepped-through hole with electrolyte entering through predrilled hole in the workpiece. (c) Cross-flow tool used to generate ribs on a surface without leaving flow lines on the part. Electrolyte is fed down one side, across the face, and up the second side. Source: Ref 2 Insulation is important to control the electric current. The insulating material should have high electrical resistivity, low (or no) water absorption, chemical resistance to the electrolyte, wear resistance, and resistance to heat at high working temperatures (200 °C, or 400 °F). It should be smooth and uniform in thickness (≥0.05 mm, or 0.002 in.). Preformed insulation can be shrink fitted with adhesives. At low electrolyte flow rates and low current density, materials like Teflon, epoxy, urethane, phenolic, and powder coating work satisfactorily. Tools with such insulation have enough life if the tool has a lip to protect the insulation from the flow force of the electrolyte. Sprayed or dipped epoxy resins are among the most effective insulating materials. Nylon, acetal, and fiberglass reinforced epoxy give better insulation.

References cited in this section 2. T.L. Lievestro, Electrochemical Machining, Machining, Vol 16, Metals Handbook, 9th ed., ASM International, 1989, p 533–541

12. A.E. DeBarr and D.A. Oliver, Ed., Electrochemical Machining, Macdonald and Co. Ltd, 1975 13. H. Tipton, The Determination of Tool Shape for Electrochemical Machining, Mach. Prod. Eng., 1968, p 325–328 14. Report No. 145, Production Engineering Research Association, Melton Mowbray, 1968 20. R.H. Nilson and Y.G. Tsuei, Free Boundary Problem for the Laplace Equation with Application to ECM Tool Design, J. Eng. Ind. (Trans. ASME), Vol 98, 1976, p 54–58 21. P. Lawrence, Computer Aided Design for ECM Electrodes, Int. J. Mach. Tools Des. Res., Vol 21, 1981, p 379–385 22. M.S. Reddy, V.K. Jain, and G.K. Lal, Tool Design for ECM: Correction Factor Method, J. Eng. Ind. (Trans. ASME), Vol 110, 1988, p 111–118 23. V.K. Jain, K. Ravi, G.K. Lal, and K.P. Rajurkar, Investigations into Tool Design for Electrochemical Drilling, Process. Adv. Mater., Vol 1 (No. 1), 1991, p 105–121 24. O.H. Narayanan, S. Hinduja, and C.F. Noble, Design of Tools for Electrochemical Machining by the Boundary Element Method, Proc. Inst. Mech. Eng., Vol 200 (C3), 1986, p 195–205 25. V.K. Jain and K.P. Rajurkar, An Integrated Approach for Tool Design in ECM, Prec. Eng., Vol 13 (No. 2), 1991, p 111–124 27. H. Tipton, Calculation of Tool Shape for ECM, Electrochemical Machining, C.L. Faust, Ed., Electrochem. Soc., 1971, p 87–102


Integrated Approach for Tool Design in ECM The present status of tool design in ECM results in low productivity, high response time to the product changes, high lead and delivery times, and high overall cost. In the future, ECM activities will be integrated by employing the capabilities of high-speed computers. This would result in integration of computer-aided design (CAD) with process planning, including computer simulation and its testing, and finally computer-aided manufacturing (CAM). An attempt for such integration would lead to enhanced productivity, higher material and equipment utilization, and better consistency and quality. When machining complicated components, it is not always possible to complete the machining process from raw material to the finished component in a single step (Ref 30). When a large amount of material is to be removed, or very high accuracy is required, the machining should be planned in two or more stages, rough and finish machining, as is done in conventional machining. The rules must be developed for the quantity of material left for finish machining so that the desired accuracy can be achieved with minimal cost (Ref 31). Sometimes machining may involve different types of ECM operations, such as die sinking, grinding, or wire cutting. Some components may need electric-discharge machining or conventional machining. The proper

sequence of these operations is important. Figure 11(a) shows high-velocity electrolyte impinging on a previously machined surface that could result in corrosion and product rejection. Figure 11(b) shows ECM in proper sequence of the same component (Ref 30).

Fig. 11 Sequence of operations. (a) Improper sequencing would result in unwanted corrosion. (b) Proper sequencing where the center was machined first. Source: Ref 30 The integrated ECM process planning (Ref 31) scheme has three components: system input, optimization, and process simulation. System input provides all the relevant data and initial information necessary to design the tool, validate the designed tool, and plan the machining of components. Optimization of machining parameters (electrolyte flow velocity, voltage, and feed rate) must be established. This can be done by multicriteria optimization (Ref 32) rather than by single-objective optimization. In multicriteria optimization of ECM process, a linear goal programming technique is employed using geometrical accuracy, material-removal rate, and tool life as objective functions. Temperature, passivation, and choking are used as constraints. Voltage, feed rate, and electrolyte flow velocity are considered design variables. According to rough or finish machining, objective functions are assigned priorities and weights. For example, in finish machining, accuracy would have first priority and metal-removal rate second. This would be reversed for rough machining. Process simulation can be used to evaluate alternative process plans and trade-offs. Process simulation is used to reduce lead time, lower cost, increase product quality, and provide a better understanding of the process (Ref 33). Simulation of the ECM process will significantly reduce and, in some cases, eliminate the iterative process of full-scale testing of tools before releasing them for actual production by the machine shop (Fig. 12).

Fig. 12 Proposed procedure for designing and testing a tool for ECM. Workpiece (W/P). Source: Ref 25 A finite-element simulator for the ECM process generates an element mesh from the stored graphic description of the part. The finite-element nodes, elements, and boundary conditions are submitted to the simulation analysis program. A postprocessor or graphic output display of the results exhibits the expected and desired workpiece profile and the designed tool profile. The process plan displays the machining parameters, electrolyte and its concentration, the code of the selected machine tool, and the names and sequence of operations. Figure 13 shows a typical ECM process-simulation plan.

Fig. 13 Simulation program for ECM. Source: Ref 25

References cited in this section 25. V.K. Jain and K.P. Rajurkar, An Integrated Approach for Tool Design in ECM, Prec. Eng., Vol 13 (No. 2), 1991, p 111–124 30. J.F. Wilson, Practice and Theory of Electrochemical Machining, Wiley Interscience, 1971

31. Gurusaran, J.L. Batra, and V.K. Jain, Computer Aided Process Planning for ECM, Tenth Int. Conf. Production Research, Nottingham, U.K., 1989, p 780–789 32. B.G. Acharya, V.K. Jain, and J.L. Batra, Multi Objective Optimization of ECM Process, Prec. Eng., Vol 8 (No. 2), 1986, p 88–96 33. K.C. Maddux and S.C. Jain, CAE for the Manufacturing Engineer: The Role of Process Simulation in Concurrent Engineering, Proc. ASME Symp. Manufacturing Simulation and Processes, A.P. Tseng, D.R. Durham, and R. Komanduri, Ed., PED 20, 1986, p 1–16


Process Capabilities and Limitations Electrochemical machining can machine highly complicated and curved shapes in a single pass. Tool life is very high so a large number of pieces can be machined by the same tool. Machinability of work materials is independent of physical and mechanical properties. Machined surfaces are stress and burr free with good surface finish. Scrap production and overall machining times are reduced. Electrochemical machining also has limitations. Accuracy of the machined component depends on tool design, degree of process control imposed, and complexity in the shapes produced. Machining of materials containing hard spots, inclusions, sand, and scale present some practical difficulties. Electrochemical machining cannot produce sharp corners and edges. The roughness of the machined surface is controlled mainly by local current density. High current density in the front gap produces low values of surface roughness. The best surface roughness varies between 0.1 and 1.0 μm (0.004 and 0.04 mil), while low value of current density in the side gap results in the high values of surface roughness (may be as rough as 5 μm or more). Variables affecting local current density ultimately affect surface roughness. The usual accuracy achievable in ECM is approximately ±0.013 mm (±0.0005 in. 0.005 in.) and ±0.25 mm (±0.010 in.) in front and side gaps, respectively. The best tolerance that can be achieved is ±0.025 mm (±0.001 in.), internal radius as 0.8 mm (0.030 in.), external radius as 0.5 mm (0.020 in.), 0.001 mm/mm (0.001 in./in.) as taper, and 0.05 mm (0.020 in.) as overcut.


Application Examples

It has been reported (Ref 12) that cost advantage of the ECM process over conventional machining varies from 4 to 1 to 9 to 1. Example 1: ECM of a Stainless Steel Nozzle. A nozzle (Fig. 14a) is made from a 316 stainless steel tube using a machine that could provide maximum current of 500 A, voltage ≤ 24 V. A copper tool was fed into one end of the tube to make conical-shaped hole. The initial current was 20 A (voltage = 17 V) and ended with the current as 310 A. Electrolyte used was NaCl (0.12 kg/L, or 1 lb/gal, of water) at 27 °C (80 °F) and 550 kPa (80 psi) pressure. Feed rate of 5.8 mm/min (0.25 in./min) took a total of 7.6 min to produce the nozzle with a surface roughness of 0.125 to 0.250 μm (5 to 10 μin.) (Ref 2).

Fig. 14 Examples of applications. (a) Finishing a conical hole in a nozzle. (b) Contouring a turbine blade surface. (c) Cutting spiral grooves in a friction plate. Source: Ref 2 Another copper tool was used to machine the radius at the other end of the hole. Parameters used were: current varied from 20 to 220 A, f = 2.5 mm/min (0.100 in./min), total machining time = 2.00 min, penetration depth = 4.95 mm (0.19 in.), and internal finish obtained = 0.125 to 0.25 μm (5 to 10 μin.) (Ref 2). Example 2: ECM of a Turbine Blade. The turbine blade (Fig. 14b) made of heat-resistant alloy (A-286 alloy) was machined by ECM that could provide maximum current of 5000 A at 2 to 20 V dc. A tool was made of

copper-tungsten alloy. The NaNO3 electrolyte (0.26 kg/L, 2.2 lb/ gal of water) at 43 °C (110 °F) was employed. Other parameters were: • • • •

Current = 100 to 170 A Pressure = 900 to 1400 kPa (130 to 205 psi) Flow rate = 7.6 L/min (2 gal/min) f = 7.6 mm/min (0.3 in./min)

The surface roughness achieved was 0.38 to 0.63 μm (15 to 25 μin.) (Ref 2). Example 3: ECM of Spiral Grooves in a Friction Disk. Seventy-two equally spaced spiral grooves (Fig. 14c), 0.38 mm (0.015 in.) deep and 1.0 mm (0.040 in.) wide were made in a friction disk (1020 steel, 60 to 75 HRB) using ECM machine, which could provide maximum current, I ≈ 600 to 650 A, voltage = 20 V, electrolyte → NaCl (0.15 kg/L, or 1.25 gal/lb of water), T = 30 to 32 °C (85 to 90 °F), P = 690 kPa (100 psi), electrolyte flow rate = 40 L/min (10 gal/min), f = 1.3 mm/min (0.050 in./min).

References cited in this section 2. T.L. Lievestro, Electrochemical Machining, Machining, Vol 16, Metals Handbook, 9th ed., ASM International, 1989, p 533–541 12. A.E. DeBarr and D.A. Oliver, Ed., Electrochemical Machining, Macdonald and Co. Ltd, 1975 V.K. Jain, Electrochemical Machining, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 143–152 Electrochemical Machining V.K. Jain, Indian Institute of Technology at Kanpur (India)

Acknowledgment The author wishes to acknowledge the suggestions of Mr. Syadwad Jain of Ohio State University, Columbus, OH, and help of Ms. Lalitha and Ms. Shaifali Mehrotra of I.I.T. Kanpur, India, in the preparation of this manuscript.


References 1. V.K. Jain, Advanced Machining Processes, Allied Publishers, Delhi, India, 2002 2. T.L. Lievestro, Electrochemical Machining, Machining, Vol 16, Metals Handbook, 9th ed., ASM International, 1989, p 533–541

3. V.K. Jain and P.C. Pandey, Investigations into the Use of Bits as a Cathode in ECM, Int. J. Mach. Tools Des. Res., Vol 22 (No. 4), 1982, p 341–352 4. J.A. McGeough, Principles of Electrochemical Machining, Chapman and Hall, London, 1974 5. H. Tipton, The Dynamics of Electrochemical Machining, Advances in Machine Tool Design and Research, Pergamon Press, Birmingham, 1964, p 509–522 6. V.K. Jain and P.C. Pandey, Tooling Design for ECM: A Finite Element Approach, J. Eng. Ind. (Trans. ASME), Vol 103, 1981, p 183–191 7. V.K. Jain and P.C. Pandey, Tooling Design for ECM, Prec. Eng., Vol 2, 1980, p 195–206 8. P.C. Pandey and H.S. Shan, Modern Machining Processes, Tata McGraw Hill, Delhi, India, 1980 9. J.F. Thorpe and R.D. Zerkle, Analytical Determination of the Equilibrium Gap in Electrochemical Machining, Int. J. Mach. Tools Des. Res., Vol 9, 1969, p 131–144 10. J. Hopenfeld and R.R. Cole, Prediction of the Dimensional Equilibrium Cutting Gap in Electrochemical Machining, J. Eng. Ind. (Trans. ASME), Vol 91, 1969, p 75–765 11. V.K. Jain, P.G. Yogindra, and S. Murugan, Prediction of Anode Profile in ECBD and ECD Operations, Int. J. Mach. Tools Manuf., Vol 27 (No. 1), 1987, p 113–134 12. A.E. DeBarr and D.A. Oliver, Ed., Electrochemical Machining, Macdonald and Co. Ltd, 1975 13. H. Tipton, The Determination of Tool Shape for Electrochemical Machining, Mach. Prod. Eng., 1968, p 325–328 14. Report No. 145, Production Engineering Research Association, Melton Mowbray, 1968 15. V.K. Jain, Vinod K. Jain, and P.C. Pandey, Corner Reproduction Accuracy in Electrochemical Drilling of Blind Holes, J. Eng. Ind. (Trans. ASME), Vol 106, 1984, p 55–61 16. W. König and H.J. Hümb, Mathematical Model for the Calculation of the Contour of the Anode in Electrochemical Machining, Ann. CIRP, Vol 25, 1977, p 83–87 17. H. Heitmann, Surface Roughness Formation and Determination of Optimum Working Conditions in Electrochemical Die Sinking, Proc. Second AIMTDR Conf., 1968, p 41–47 18. Y.G. Tsuei, C.H. Yen, and R.H. Nilson, Theoretical and Experimental Study of Workpiece Geometry in Electrochemical Machining, ASME, Paper 76-WA/Prod, presented in WAM, 1977, p 1–5 19. O.H. Narayanan, S. Hinduja, and C.F. Noble, The Prediction of Workpiece Shape During Electrochemical Machining by the Boundary Element Method, Int. J. Mach. Tools Des. Res., Vol 26 (No. 3), 1986, p 323–338 20. R.H. Nilson and Y.G. Tsuei, Free Boundary Problem for the Laplace Equation with Application to ECM Tool Design, J. Eng. Ind. (Trans. ASME), Vol 98, 1976, p 54–58 21. P. Lawrence, Computer Aided Design for ECM Electrodes, Int. J. Mach. Tools Des. Res., Vol 21, 1981, p 379–385

22. M.S. Reddy, V.K. Jain, and G.K. Lal, Tool Design for ECM: Correction Factor Method, J. Eng. Ind. (Trans. ASME), Vol 110, 1988, p 111–118 23. V.K. Jain, K. Ravi, G.K. Lal, and K.P. Rajurkar, Investigations into Tool Design for Electrochemical Drilling, Process. Adv. Mater., Vol 1 (No. 1), 1991, p 105–121 24. O.H. Narayanan, S. Hinduja, and C.F. Noble, Design of Tools for Electrochemical Machining by the Boundary Element Method, Proc. Inst. Mech. Eng., Vol 200 (C3), 1986, p 195–205 25. V.K. Jain and K.P. Rajurkar, An Integrated Approach for Tool Design in ECM, Prec. Eng., Vol 13 (No. 2), 1991, p 111–124 26. M.B. Nanyakkara and C.N. Larsson, Computation and Verification of Workpiece Shape in Electrochemical Machining, 20th Int. Machine Tool Design and Research Conf., UMIST, U.K., 1976, p 617–624 27. H. Tipton, Calculation of Tool Shape for ECM, Electrochemical Machining, C.L. Faust, Ed., Electrochem. Soc., 1971, p 87–102 28. V.K. Jain and P.C. Pandey, On the Reproduction of Corner Anode Profile During Electrochemical Drilling of Blind Holes, J. Eng. Ind. (Trans. ASME), Vol 106 (No. 1), 1984, p 55–61 29. W. König and D. Pahl, Accuracy and Optimal Working Conditions in ECM, Ann. CIRP, Vol 8 (No. 2), 1970, p 223–230 30. J.F. Wilson, Practice and Theory of Electrochemical Machining, Wiley Interscience, 1971 31. Gurusaran, J.L. Batra, and V.K. Jain, Computer Aided Process Planning for ECM, Tenth Int. Conf. Production Research, Nottingham, U.K., 1989, p 780–789 32. B.G. Acharya, V.K. Jain, and J.L. Batra, Multi Objective Optimization of ECM Process, Prec. Eng., Vol 8 (No. 2), 1986, p 88–96 33. K.C. Maddux and S.C. Jain, CAE for the Manufacturing Engineer: The Role of Process Simulation in Concurrent Engineering, Proc. ASME Symp. Manufacturing Simulation and Processes, A.P. Tseng, D.R. Durham, and R. Komanduri, Ed., PED 20, 1986, p 1–16 34. J.L. Bennett, Building Decision Support System, Addison Wesley, 1983 35. L. Kops, Effect of Pattern of Grain Boundary Network on Metal Removal Rate in Electrochemical Machining Process, J. Eng. Ind. (Trans. ASME), Vol 98, 1976, p 360–368 V.K. Jain, Electrochemical Machining, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 143–152 Electrochemical Machining V.K. Jain, Indian Institute of Technology at Kanpur (India)


G.F. Benedict, Nontraditional Manufacturing Processes, Marcel Dekker, 1987

• • • •

• • • • • • • • •

D.T. Chin, Anodic Film and ECM Dimensional Control: A Study of Steel Electrodes in Solutions Containing Na2SO4 and NaClO4, J. Electrochem. Soc., Vol 121 (No. 12), 1974, p 1592–1595 W.G. Clark and J.A. McGeough, Temperature Distribution in the Gap in the Electrochemical Machining, J. Appl. Electrochem., Vol 7, 1977, p 277–286 A.K.M. DeSilva and J.A. McGough, Computer Application in Unconventional Machining, J. Mater. Process. Technol., Vol 107, 2000, p 276–282 J. Kozak, M. Chuchro, A. Ruszaj, and K. Karbowski, The Computer Aided Simulation of Electrochemical Process with Universal Spherical Electrodes When Machining Sculptured Surfaces, J. Mater. Process. Technol., Vol 107, 2000, p 283–287 J. Kozak, K.P. Rajkumar, and R. Balkrishna, Study of Electrochemical Jet Machining Process, J. Manuf. Sci. Eng. (Trans. ASME), Vol 1996, p 490–498 M.E. Merchant, Newer Methods for the Precision Working of Metals: Research and Present Status, Proc. Int. Conf. on Production Research (ASME), 1962, p 93–107 K.P. Rajurkar, C.L. Schnacker, and R.P. Lindsay, Some Aspects of ECM Performance, Ann. CIRP, Vol 37 (No. 1), 1988, p 183–186 K.P. Rajurkar and D. Zhu, Improvement of Electrochemical Machining Accuracy by Using Orbital Electrode Movement, Ann. CIRP, Vol 48 (No. 1), 1999, p 139–142 A. Ruszaj, Investgation Aiming to Increase Electrochemical Machining Accuracy, Prec. Eng., Vol 12 (No. 1), 1999, p 43–48 K. Seimiya, An Approximate Method for Predicting Gap Profile in Electrochemical Machining, Proc. Fifth Int. Conf. Prod. Eng. (Tokyo, Japan), 1984, p 389–394 S.M. Trendler, ECM and Ultrasonic Machining for Faster Die Machining, Die Cast. Eng., Jan/Feb 1985, p 66–67 F. Zawistowski, New System of Electrochemical Form Machining Using Universal Rotating Tool, Int. J. Mach. Tools Manuf., Vol 30 (No. 3), 1990, p 475–483 M. Zybura Skrabalak and A. Ruszaj, The Influence of Electrode Surface Geometrical Surface on Electrochemical Soothing Process, J. Mater. Process. Technol., Vol 107, 2000, p 288–292

V.K. Jain, Electrochemical Allied Processes, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 153–156

Electrochemical Allied Processes V.K. Jain, Indian Institute of Technology Kanpur

Introduction SPECIFIC MACHINING PROCESSES that employ electrochemical machining (ECM) technology include deburring, and deep-hole drilling. The principles of electrochemical machining are presented in the article “Electrochemical Machining” in this Volume. This article shows the applications of ECM to specific processes and discusses a variation of the steady-state process, pulse machining.

V.K. Jain, Electrochemical Allied Processes, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 153–156 Electrochemical Allied Processes V.K. Jain, Indian Institute of Technology Kanpur

Electrochemical Deburring A designer usually considers aspects of a component such as material, form, dimensional accuracy, surface texture, and heat treatment, but perhaps not the surface integrity and edge quality (Ref 1). These last two factors are important aspects for the performance and life of a product, however. When a component is processed by a conventional machining method, it is usually left with burrs along intersecting surfaces. Such burrs are unwanted and could be a hazard when handling the component, but they can be removed by one of several deburring processes. In modern industrial technology, the deburring process has attained considerable importance because of stringent quality standards. Presence of burrs on a component affects its functional, physiological, and aesthetic requirements (Ref 2). Burrs create a hindrance in assembly, they may degrade corrosion preventive or aesthetic coatings, and loosened burrs may result in excessive wear of the sliding or rolling surfaces. Different types of burrs are listed in Table 1 (Ref 3). Table 1 Types of burrs formed during different manufacturing methods Type of burr Compressive burr

Schematic

Remarks The burr produced in blanking and piercing operations in which slug separates from the parent material under compressive stress

Cutting-off burr

A projection of material left when the workpiece falls from the stock

Corner burr

Intersection of three or more surfaces

Edge burr

Intersection of two surfaces

Entrance burr

Cutting tool enters in the workpiece

Exit burr Feather burr

Cutting tool exits the workpiece Fine or thin burr

Flash burr

Portion of flash remaining on the part after trimming

Hanging burr

Loose burr not firmly attached to the workpiece

Rollover burr

Burr formed when it exits over a surface and allows the chips to be rolled away

Source: Ref 3 Deburring processes can be mechanical, abrasive, thermal, chemical, and electrochemical. Mechanical deburring using brushes or scrapers is unreliable and is a burr-minimizing process, which does not meet the requirements of high edge quality. The abrasive deburring processes (tumbling, sand blasting, or vibratory) have low reliability, poor uniformity, low metal removal rate, and tend to charge the workpiece with grit. In thermal deburring, the deburring chamber temperature is about 3500 °C (6330 °F), which burns out burrs and sharp edges on the component. In chemical deburring, burrs are dissolved in a chemical medium. In electrochemical deburring (ECD), the principle of anodic dissolution is applied to dissolve the burrs. Electrochemical deburring works on the same principle as ECM. The tool is either insulated on all surfaces except the part that is adjacent to the burrs, or a bit type of tool is used (Ref 4). The tool tip (or bit) should overlap the area to be deburred by about 1.5 to 2.0 mm (0.06 to 0.08 in.). Electrochemical deburring is useful for burrs located in inaccessible areas where other deburring processes are not effective (Ref 5, 6). In ECD, the magnitude of current and electrolyte flow rate are lower than ECM. Secondly, the tool is stationary (Fig. 1). The current magnitude and its duration of flow to suit a particular component are determined by trial. The commonly used electrolytes are NaCl and NaNO3. The interelectrode gap (IEG) is usually in the range of 0.1 to 0.3 mm (0.004 to 0.012 in.). A deburred steel component shows a localized deposit (Fe3O4— dark gray and ≤1 μm, or 4 × 10-6 in., thick) that is a reaction product of the process. This deposit is conductive and disappears during heat treatment of the components.

Fig. 1 Electrochemical deburring at the intersection of two holes Machine tools for ECD are usually designed with multiple work stations served from a single power supply. Their elements and functions are the same as those of ECM tools except for the tool-feed system. Before applying ECD for a particular type of job, the thickness, shape, and repeatability of the burrs on the parts must be known. The more uniform the shape and size of the burrs, the more efficient the burr-removal operation. Further, the tool head that dissolves the burrs should be shaped as a replica of the contour of the work. Portable ECD units equipped with 50 A power supplies (Ref 4) are also available and are useful to deburr tubes and pipes of varying configuration, length, and cross- sectional area. Applications in consumer appliances, biomedical, aerospace, and automobile industries include gears, splines, drilled holes, milled components, fuel supply and hydraulic system components, and intersecting holes in crankshafts (Ref 2). Apart from economics, ECD gives higher reliability, reduced operation time, more uniformity, and it can be easily automated. Deburring components takes about 45 to 60 s cycle time on a 500 A machine. The hydroxides

removed during ECD can be safely used as a raw material for the lapping paste. The components should be thoroughly washed before deburring, and properly cleaned after deburring (Ref 2).

References cited in this section 1. K. Takazawa, The Challenge of Burr Technology and Its Worldwide Trends, Bull. JSPE, Vol 22 (No. 3), 1988, p 165–170 2. M.G.J. Naidu, Electrochemical Deburring for Quality Production, Proc. Winter School on AMT, V.K. Jain and S.K. Choudhury, Ed., IIT Kanpur (India), 1991 3. V.K. Jain, Advanced Machining Processes, Allied Publishers, New Delhi, 2002 4. V.K. Jain and P.C. Pandey, Investigations into the Use of Bits as a Cathode in ECM, Int. J. Mach. Tools Des. Res., Vol 22 (No. 4), 1982, p 341–352 5. G.F. Benedict, Non-traditional Manufacturing Processes, Marcel Dekker, 1987 6. E. Rumyantsev and A. Davydev, Electrochemical Machining of Metals, Mir Publishers, Moscow, 1989

V.K. Jain, Electrochemical Allied Processes, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 153–156 Electrochemical Allied Processes V.K. Jain, Indian Institute of Technology Kanpur

Electrochemical Deep-Hole Drilling Deep small holes are required in many applications, such as turbine rotor and stator assembly for cooling, crankshaft oil holes, holes in fuel- injection nozzles, and holes in spinnerets. The principle of anodic dissolution has been applied for making such small-diameter deep holes in hard-to-machine materials. This process is also known as shaped-tube electrolytic machining (Ref 7). This process basically differs from conventional ECM in two aspects: either a mixture (neutral salt and acidic electrolyte) or acidic electrolyte is used, and a coated tubular cathode is used. The mixture or acidic electrolyte is added to circumvent the production of insoluble precipitate (sludge) during conventional ECM, which inhibits machining process by obstructing the flow path of the electrolyte. The coated tubular cathode controls the magnitude of overcut. Certain minimum overcut is desirable to serve as a return path for the electrolyte (Ref 8). If acid (HCl, H2SO4, or HNO3) is used as electrolyte, then periodic reversal of the polarity for a very short duration of time (5–10 s) is essential to remove any film built up on the cathode. Such reversal is usually not required if a mixture of the electrolytes is used (Ref 7, 8, 9). Simultaneous drilling of many holes is possible (Ref 7, 9). A modern air-pressure turbine rotor and stator assembly may have 20,000 cooling holes having diameters ranging from 1 to 4 mm (0.04 to 0.16 in.) and aspect ratios of 40 to 200. The rotor blades are subjected to high stresses, very high temperature and vibration, and are made of difficult-tomachine superalloys. Analysis of the data in Table 2 (Ref 8) indicates that laser-beam machining and electronbeam machining cannot be used because of upper limit of thickness of the workpiece (only 18 mm, or 0.7 in., or so) that can be drilled. Electrical discharge machining (EDM) suffers in comparison with electrochemical deephole drilling due to relatively lower aspect ratio, shallower depths, poorer surface finish, and surface integrity. Table 2 Comparison of process capabilities of the advanced methods of deep-hole drilling

Parameter

Process EDM Hole diameter, 0.125–6.25 (0.005–0.25) mm (in.) Hole depth, mm (in.) 3.125 (0.123) Common Max 62.5 (2.5) Ultimate Aspect ratio 10:1 Typical 20:1 Maximum Cutting rate, 0.0125 (0.0005) mm/s (in./s) Finish, μm (μin.) 1.6–3.2 (63–125) Operating voltage Surface integrity

30–100 V

STEM ESD 0.75–2.5 (0.03– 0.125–0.875 0.1) (0.005–0.034)

LBM 0.125–1.25 (0.005–0.5)

EBM 0.025–1.0 (0.001–0.04)

125 (4.9) 900 (35.4)

18.75 (0.74) 25 (1.0)

5.0 (0.2) 17.5 (0.7)

2.5 (0.1) 7.5 (0.3)

16:1 300:1 0.025 (0.001)

16:1 40:1 0.025 (0.001)

16:1 75:1 150 mW/cm2. During the 1980s, both the performance and endurance of MCFC stacks showed dramatic improvements. Several MCFC stack developers have produced cell stacks with cell areas up to 1 m2 (10.8 ft2) cells. Tall, full-scale stacks fabricated to date include several FuelCell Energy (FCE, formerly Energy Research Corporation) 300 plus cell stacks with ~9000 cm2 (1395 in.2) cell area producing >250 kW. Table 4 Evolution of cell-component technology for molten-carbonate fuel cells Component Anode

Circa 1965 Pt, Pd, or Ni

Circa 1975 Ni-10Cr

Current status Ni-Cr/Ni-Al/Ni-Al-Cr 3–6 μm (0.12–0.24 mil) pore size 45–70% initial porosity 0.20–1.5 mm (0.008–0.06 in.) thickness

Cathode

Ag2O or lithiated NiO

Lithiated NiO

0.1–1 m2/g Lithiated NiO-MgO 7–15 μm (0.3–0.6 mil) pore size

70–80% initial porosity 60–65% after lithiation and oxidation 0.5–1 mm (0.02–0.04 in.) thickness

Electrolyte support

MgO

Mixture of α-, β-, and γLiAlO2

0.5 m2/g γLiAlO2, αLiAlO2 0.1–12 m2/g

2

10–20 m /g Electrolyte(a), wt%

52Li-48Na

1.8 mm (0.07 in.) thickness 62Li-38K

0.5–1 mm (10.02–0.04 in.) thickness 62Li-38K

43.5Li-31.5Na-25K

Hot-press “tile”

60Li-40Na

“Paste”

1.8 mm (0.07 in.) thickness

51Li-48Na Tape cast

0.5–1 mm (0.02–0.04 in.) thickness (a) A mole percent of alkali carbonate salt The conventional process used to fabricate electrolyte structures until about 1980 involved hot pressing (~34 MPa, or 5 ksi) mixtures of LiAlO2 and alkali carbonates (typically >50 vol% in liquid state) at temperatures slightly below the melting point of the carbonate salts (e.g., 490 °C, or 915 °F, for electrolyte containing 62 mol% Li2CO3-38 mol% K2CO3). These electrolyte structures also called electrolyte tiles, were relatively thick (1 to 2 mm, or 0.04 to 0.08 in.) and difficult to produce in large sizes because large tooling and presses were required. The electrolyte structures produced by hot pressing are often characterized by void spaces ( kp. For type III of EEO, the erosion causes spalling of the oxide scale, as seen in Fig. 5. The regime of oxidation-modified erosion, where the oxidation enhances the ability of the erodent to remove material. In this case, the total removal is a function of time. Bulk erosion, where the erodent energy is able to cut through the surface scale and remove the bulk material below the scale. In this case, the total removal is also just a function of time. Maps of these regimes have been produced to help illustrate the different mechanisms in operation (Ref 28).

Fig. 7 Erosion-corrosion interaction regimes, in order of increasing corrosion. Source: Ref 27 In a study of stainless steels exposed to a combustion atmosphere from a propane-fired abrasive-jet machine at 975 °C, (1790 °F), Ives (Ref 29) found the velocity dependence of erosion rate, E, for a 90° impact of ν1.8. The tests were conducted using a commercial grade of SiC with the average particle size of ~135 μm. By comparison, the velocity dependence at room temperature was ν2.5. The results show the thickness loss decreased with increasing chromium content (Table 1). Aluminum additions had a beneficial effect under the conditions of excess fuel at 1000 °C (1800 °F) and a particle velocity of 55 m/s (180 ft/s, or 123 mph). Table 1 Erosion-corrosion test results, using excess fuel at 1000 °C (1800 °F) and 55 m/s (123 mph) particle velocity of ~135 μm (5 mils) SiC abrasive

Alloy Major alloying constituents, wt% Thickness loss, μm/h α = 90° α = 45° 19 31 I-671 51Cr, 47Ni 29 62 I-601 22Cr, 61Ni, 1.2Al 70 80 I-800 33Cr, 21Ni, 42Fe 25Cr, 21Ni, 52Fe 73 80 310 25Cr, 72Fe 90 130 446 17Cr, 12Ni, 68Fe, 2.5Mo 150 160 316 19Cr, 10Ni, 68Fe 195 190 304 α is angle of impingement. Source: Ref 29 Erosion-corrosion maps have been developed in order to show the effects of different variables and to identify different degradation mechanisms (Ref 30). These maps should help the design engineer select materials and understand the mechanisms involved in erosion-corrosion at elevated temperatures (Ref 31). Fretting Wear. Fretting is a form of wear that occurs between contacting surfaces subjected to cyclic relative motion of extremely small amplitude, usually in the range of 1 to 100 μm. This wear typically occurs on surfaces that are not intended to move in this manner, such as bearing mounts, hubs, and shafts. This is often considered a form of corrosion since usually the debris formed rapidly oxidizes. The relative motion between the two surfaces causes fretting. The contact creates an abrasive wear on the oxide films on both surfaces, removing the passive layer while leaving oxide debris particles. These debris particles contribute to the wear process, increasing the wear as time passes. The passive layer on a fretting part is continuously restored and destroyed as new material is exposed to the atmosphere and oxidized. The destruction of the layer produces additional wear debris, which goes on to produce more corrosion and wear. In steels, this debris layer can form a white etch-resistant layer, which is extremely hard and brittle (Fig. 8). This layer has been described as martensite.

Fig. 8 Cross section of a white layer of martensite produced by fretting, after 105 fretting cycles and a contact stress of 40 MPa (5.8 ksi). Source: Ref 32 While not usually considered a major issue, fretting is known to occur at elevated temperatures. Examples of this are found in gas and steam turbines. Fretting at elevated temperatures can form low friction glazes (Ref 33), which will reduce the loss material. Sliding Wear. Adhesive, oxidational, and sliding wear refers to the wear that takes place when one surface “slides” over another. This typically occurs in machining, unlubricated machinery, and material-handling operations. Machining is an example of the adhesive-wear process. At the tip of the cutting tool, the chip is separated from the bulk of the part. This chip may strike the tool further back. This can lead to both the dulling of the tip and crater wear further back. During high-speed machining tool temperatures can reach 1000 °C (1800 °F), which can cause oxidation of the tool surface. Typically, the self- formed oxide layer will be less resistant to adhesive wear than the bulk of the tool. As a consequence, wear damage accelerates as competition between the removal of the oxide film and the formation of fresh oxide occurs.

Oxidation has long been considered a fundamental aspect of sliding wear (Ref 34). Oxidation can occur at relatively low sliding velocities, in the range of 0.2 to 1 m/s (8 to 40 in./s) (Ref 35). Oxidation accelerates with increasing temperature and can lead to the formation of a protective layer under many conditions. These layers tend to be in the range of 3 to 4 μm thick and are quite beneficial in unlubricated conditions.

References cited in this section 8. R. Blickensderfer, B.W. Madsen, and J.H. Tylczak, Comparison of Several Types of Abrasive Wear Tests, Proc. Int. Conf. Wear of Materials (Vancouver, B.C., Canada), American Society of Mechanical Engineers, 1985, p 313 9. A.J. Sedriks and T.O. Mulhearn, The Effect of Work-Hardening on the Mechanics of Cutting in Simulated Abrasive Processes, Wear, Vol 7, 1964, p 451–459 10. T. Sasaki and K. Okamura, The Cutting Mechanism of Abrasive Grain, Bull. JSME, Vol 12, 1960, p 547 11. K. Kato, Wear Mode Transitions, Scr. Metall., Vol 24, 1990, p 815–820 12. J.F. Archard, Contact and Rubbing of Flat Surfaces, J. Appl. Phys., Vol 24, 1953, p 981 13. G. Sundararajan, A New Model for Two- Body Abrasive Wear Based on the Localization of Plastic Deformation, Wear, Vol 117, 1987, p 1–35 14. M.M. Khruschov, Resistance of Metals to Wear by Abrasion, as Related to Hardness, Proc. Conf. Lubrication Wear, Institution of Mechanical Engineers, London, 1957, p 655–659 15. S. Soemantri, A.C. McGee, and I. Finnie, Some Aspects of Abrasive Wear at Elevated Temperatures, Wear of Materials, 1985, Proc. Int. Conf. Wear of Materials, American Society of Mechanical Engineers, 1985, p 338–344 16. H. Berns and B Wewers, Development of an Abrasion Resistant Steel Composite with in situ TiC Particles, Wear, Vol 251, 2001, p 1386–1395 17. G.P. Tilly, Erosion Caused by Impact of Solid Particles in Wear, Erosion, Vol 13, Treatise of Materials Science and Technology, Academic Press, 1979 18. A.W. Ruff and S.M. Wielerhorn, Erosion by Solid Particle Impact, Erosion, Vol 16, Treatise on Material Science and Technology, Academic Press, 1979 19. W.F. Adler, Ed., STP 664, Erosion: Prevention and Useful Applications, ASTM, 1979 20. J.E. Ritter, Ed., Erosion of Ceramic Materials, Trans. Tech. Publications, Zurich, Switzerland, 1992 21. T.H. Kosel, Solid Particle Erosion, Friction, Lubrication, and Wear Technology, Vol 18, ASM Handbook, ASM International 1992, p 199–213 22. I. Finnie, Some Reflections on the Past and Future of Erosion, Wear, Vol 186–187, 1995, p 1–10 23. G. Sundararajan, The Solid Particle Erosion of Metallic Materials: The Rationalization of the Influence of Material Variables, Wear, Vol 186–187, 1995, p 129–144 24. I. Hutchings, and J. Little, Proc. Eighth Int. Conf. Erosion by Liquid and Solid Impact: ELSI VIII, Wear, Vol 186–187, 1995

25. J. Little, Proc. Ninth Int. Conf. Erosion by Liquid and Solid Impact: ICEAW I, Wear, Vol 233–235, 1999 26. R.H. Barkalow and F.S. Pettit, Corrosion/ Erosion of Materials in Coal Combustion Gas Turbines, Proc. Conf. Corrosion/Erosion of Cal Conversion System Materials (Berkeley, CA), 24–26 Jan 1979, National Association of Corrosion Engineers, 1979, p 139–173 27. D.M. Rishel, F.S. Petit, and N. Birks, Some Principal Mechanisms in the Simultaneous Erosion and Corrosion Attack of Metals at High Temperature, paper 16, Proc. Conf. Corrosion-Erosion-Wear of Materials at Elevated Temperatures (Berkeley, CA), 31 Jan–2 Feb 1990 A.V. Levy, Ed., NACE/ EPRI/LBL/DOE-FE 28. D.J. Stephenson and J.R. Nicholls, Modelling the Influence of Surface Oxidation on High Temperature Erosion, Wear, Vol 186–187, 1995, p 284–290 29. L.K. Ives, Erosion of 310 Stainless Steel at 975 °C in Combustion Gas Atmospheres, Trans. ASME, April 1977, p 126–132 30. M.M. Stack and L. Bray, Interpretation of Wastage Mechanisms of Materials Exposed to Elevated Temperature Erosion-Corrosion Using Erosion-Corrosion Maps and Computer Graphics, Wear, Vol 186–187, 1995, p 273–283 31. M.M. Stack, Looking Beyond the Millennium: Critical Issues in the Evaluation of Materials Performance for Resistance to Erosive Wear in Corrosive Conditions, Wear, Vol 233–235, 1999, p 484– 496 32. J. Dobromirski and I.O. Smith, Metallographic Aspects of Surface Damage, Surface Temperature, and Crack Initiation in Fretting Fatigue, Wear, Vol 117, 1987, p 347–357 33. M.M. Hamdy and R.B. Waterhouse, The Fretting-Fatigue Behavior of a Nickel- Based Alloy (Inconel 718) at Elevated Temperatures, Proc Int. Conf Wear of Materials 1979, American Society of Mechanical Engineers, 1979, p 351–355 34. T.F.J. Quinn, J.L. Sullivan, and D.M. Rowson, Origins and Development of Oxidational Wear at Low Ambient Temperatures, Wear, Vol 94, 1984, p 175–191 35. G. Straffelini, D. Trabucco, and A. Molinari, Oxidative Wear of Heat-Treated Steels, Wear, Vol 250, 2001, p 485–491

J.H. Tylczak and T.A. Adler, Gaseous Corrosion-Wear Interactions, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 338–344 Gaseous Corrosion-Wear Interactions Joseph H. Tylczak and Thomas A. Adler, Albany Research Center, U.S. Department of Energy

Measurement The techniques of measuring wear and corrosion have resulted in the creation of many standard practices and specialized pieces of equipment. This Volume outlines many of the corrosion-measurement techniques, while

Friction, Wear, and Lubrication, Vol 18, ASM Handbook, (Ref 36) outlines many of the wear-measurement techniques. Jet, whirling arm, and slinger tests have all been used successfully in the field of erosion. While the jet test (Ref 37) is most commonly used, there are definite advantages to the whirling-arm test. With the whirling-arm test, the impact velocity is well known, and an entire face of the specimen can be eroded, producing a more uniform surface. The machining test is the most commonly used high-temperature abrasive test, because the process of machining produces an elevated temperature. For slow abrasive elevatedtemperature tests one of the most commonly used is a high- temperature ring-on-disk test. Measurement of the interaction between corrosion and wear modes of damage is more difficult. One standard for measuring these interactions is ASTM G 119 (Ref 38). The scope of this standard is aqueous corrosion and wear, and applies to systems in liquid solutions or slurries. However, aspects of it can be adapted to dry corrosion and wear interactions as well. The measurement of corrosion, wear, and corrosion/wear interactions is a multistep process. Each component of the interaction must be measured separately. The results may then be combined to create a complete picture of the damage process. An example of this process is the measurement of the degradation on combustion walls in coal-fired boilers. First, one must fully understand the environment around the combustion walls. This would include knowing the temperature, gas chemistry, velocity of the gases, and the abrasive causing the erosion. After determining these details, the following tests might be used for measurements. First, an oxidation test should be performed in the combustion gas in the boiler. In this case, the gas should include the impurities (such as SO2 and HCl) as well as the major gases, since they will greatly affect the scales formed. This measurement will show the loss rate due to the gaseous corrosion or oxidation process. An erosion test would be run, using an abrasive similar in both type and size to that found in the ash of the boiler. In the case of a coal-fired boiler, a quartz abrasive would probably best represent the abrasive found in the ash. The erosion test would need to be run in a neutral (nonoxidizing) atmosphere to block the effects of the corrosion. Elevated temperature would be necessary, since temperature alone has been shown to affect erosion rates. Different test velocities for the abrasive may be run to determine the sensitivity of the erosion rate to velocity. Finally, combined wear-corrosion tests would need to be run, fully simulating the boiler conditions. The synergistic effect can then be determined (after converting all the loss rates to a common unit such as surface recession rate in mm/day) by subtracting the individual corrosion and erosion results from the final combined test. The resulting number represents the degree to which oxidation-wear interactions increase the wear: Ws = WT - W e - C

(Eq 9)

where Ws is the erosion-corrosion synergy or interaction term, WT is the total loss, We is the wear due to erosion, and C is the loss due to corrosion. The loss from erosion-corrosion interaction can then be divided by the total loss to get the percentage loss due to the interaction (Se-c): Se-c = Ws/WT

(Eq 10)

References cited in this section 36. P.J. Blau, Ed., Friction, Lubrication, and Wear Technology, Vol 18, ASM Handbook, ASM International, 1992 37. “Standard Test Method for Conducting Erosion Tests by Solid Particle Impingement Using Gas Jets,” G 76-02, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM, 2002 38. “Standard Guide for Determining Synergism Between Wear and Corrosion,” G 119- 93, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM, 1998


Control There is no definitive answer to the problem of wear, corrosion, and wear-corrosion interactions. In some environments, a trade-off between corrosion and wear can lead to a minimum of material loss. At intermediate temperatures, a corrosive layer can serve to protect the bulk material from wear (Ref 39). At high temperatures, over 800 °C (1500 °F), protective mixed-oxide layers can form, causing a decrease in the total loss rate. Changing temperatures in manufacturing environments may have the potential to decrease material loss. Some abrasives grow softer with increased temperatures and thus decrease wear. Additional research needs to be done to determine if there are any manufacturing processes where this would be a viable solution. In certain circumstances, changing the physical environment may offer potential for decreased wear and corrosion. Sulfur reacts with metals to form nonprotective coatings. Decreasing the sulfur compounds in a manufacturing environment, such as using a lower-sulfur coal, can thus reduce the material loss associated with both wear and corrosion. Efforts have been made to identify advanced materials resistant to wear-corrosion (Ref 40). Coatings, for example, have been shown to offer viable wear-corrosion protection. This enables one to put high-alloy materials on the surface where they are most needed. In some cases, applying coatings to the base material can minimize material loss. There are six common coating techniques used for wear-corrosion control. Physical vapor disposition (PVD) produces a very thin (2–10 μm, or 0.08–0.4 mil) protective layer of a material that is vaporized and deposited. Chemical vapor deposition (CVD), which produces a 5 to 15 μm (0.2 to 0.6 mil) protective layer, is made by a gas-phase decomposition of a material onto a part. Flame spray uses a heated flame to partially melt a thin (0.25–0.6 mm, or 0.01–0.02 in.) coating, commonly a high- chrome alloy, and adhere it to the base material. Diffusion coating involves covering the surface to be treated with a mostly inert material such as alumina, and heating resulting in an interaction that creates a thin (0.25–0.3 mm, or 0.01–0.012 in.) protective layer. Weld overlays consist of a thicker (1–2 mm, or 0.04–0.08 in.) protective layer of higheralloy metal welded over the base material. Thick (12–20 mm, or 0.47–0.79 in.) refractory coatings, such as silicon carbide base materials, can be bonded to a metallic base. Experience indicates that these all have useful applications, but overlays and refractory coatings may be the most resistant to severe erosion-corrosion. Different environments require different solutions. In some instances, engineering changes to temperatures, equipment design, flow rates, and so forth, can minimize wear and corrosion. Lubrication, when it can be used, is a partial solution. Selecting advanced materials, such as ceramics or intermetallics, is another possible option. There are three possibilities for reducing erosion-corrosion damage: (a) modify the erodent, (b) change the material, and (c) change the aerodynamics so the particle impacts are less damaging (Ref 41). The more that is known about wear-corrosion interaction and the environmental specifics of a given manufacturing process, the more likely it is that solutions can be found to the wear-corrosion problem.

References cited in this section 39. A. Levy, J. Yan, and J. Patterson, Elevated Temperature Erosion of Steels, Wear, Vol 108, 1986, p 43– 60 40. V. Shanov and W. Tabakoff, Erosion Resistance of Coatings for Metal Protection at Elevated Temperatures, J. Surf. Coat. Technol., Vol 86–87, 1997, p 86–93 41. J. Stringer, Practical Experience with Wastage at Elevated Temperatures in Coal Combustion Systems, Wear, Vol 186–187, 1995, p 11–27


References 1. S. Collins, Biomass-Fired Plants Face Down Near-Term Challenges, Power, Vol 137, 1993, p 51–53 2. M.A. Uusitalo, P.M.J. Vuoristo, and M.A. Mäntylä, Elevated Temperature Erosion- Corrosion of Thermal Sprayed Coatings in Chlorine Containing Environments, Wear, Vol 252, 2002, p 586–594 3. B.S. Mann and B. Prakash, High Temperature Friction and Wear Characteristics of Various Coating Materials for Steam Valve Spindle Application, Wear, Vol 240, 2000, p 223–230 4. P.M. Rogers, I.M. Hutchings, and J.A. Little, Coatings and Surface Treatments for Protection against Low-Velocity Erosion- Corrosion in Fluidized Beds, Wear, Vol 186–187, 1995, p 238–246 5. H. Berns and S. Koch, High Temperature Sliding Abrasion of a Nickel-Base Alloy and Composite, Wear, Vol 225–229, 1999, p 154–162 6. S. Dolinšek, B. Šuštaršič, and J. Kopač, Wear Mechanisms of Cutting Tools in High- Speed Cutting Processes, Wear, Vol 250, 2001, p 349–356 7. H. Stewart, R. Rapp, M. Padilla, and A. Ried, Proc Tenth Int. Wood Machining Seminar, Forest Products Laboratory, University of California, Berkeley, CA, 1985, p 54–71 8. R. Blickensderfer, B.W. Madsen, and J.H. Tylczak, Comparison of Several Types of Abrasive Wear Tests, Proc. Int. Conf. Wear of Materials (Vancouver, B.C., Canada), American Society of Mechanical Engineers, 1985, p 313 9. A.J. Sedriks and T.O. Mulhearn, The Effect of Work-Hardening on the Mechanics of Cutting in Simulated Abrasive Processes, Wear, Vol 7, 1964, p 451–459 10. T. Sasaki and K. Okamura, The Cutting Mechanism of Abrasive Grain, Bull. JSME, Vol 12, 1960, p 547 11. K. Kato, Wear Mode Transitions, Scr. Metall., Vol 24, 1990, p 815–820 12. J.F. Archard, Contact and Rubbing of Flat Surfaces, J. Appl. Phys., Vol 24, 1953, p 981 13. G. Sundararajan, A New Model for Two- Body Abrasive Wear Based on the Localization of Plastic Deformation, Wear, Vol 117, 1987, p 1–35 14. M.M. Khruschov, Resistance of Metals to Wear by Abrasion, as Related to Hardness, Proc. Conf. Lubrication Wear, Institution of Mechanical Engineers, London, 1957, p 655–659 15. S. Soemantri, A.C. McGee, and I. Finnie, Some Aspects of Abrasive Wear at Elevated Temperatures, Wear of Materials, 1985, Proc. Int. Conf. Wear of Materials, American Society of Mechanical Engineers, 1985, p 338–344

16. H. Berns and B Wewers, Development of an Abrasion Resistant Steel Composite with in situ TiC Particles, Wear, Vol 251, 2001, p 1386–1395 17. G.P. Tilly, Erosion Caused by Impact of Solid Particles in Wear, Erosion, Vol 13, Treatise of Materials Science and Technology, Academic Press, 1979 18. A.W. Ruff and S.M. Wielerhorn, Erosion by Solid Particle Impact, Erosion, Vol 16, Treatise on Material Science and Technology, Academic Press, 1979 19. W.F. Adler, Ed., STP 664, Erosion: Prevention and Useful Applications, ASTM, 1979 20. J.E. Ritter, Ed., Erosion of Ceramic Materials, Trans. Tech. Publications, Zurich, Switzerland, 1992 21. T.H. Kosel, Solid Particle Erosion, Friction, Lubrication, and Wear Technology, Vol 18, ASM Handbook, ASM International 1992, p 199–213 22. I. Finnie, Some Reflections on the Past and Future of Erosion, Wear, Vol 186–187, 1995, p 1–10 23. G. Sundararajan, The Solid Particle Erosion of Metallic Materials: The Rationalization of the Influence of Material Variables, Wear, Vol 186–187, 1995, p 129–144 24. I. Hutchings, and J. Little, Proc. Eighth Int. Conf. Erosion by Liquid and Solid Impact: ELSI VIII, Wear, Vol 186–187, 1995 25. J. Little, Proc. Ninth Int. Conf. Erosion by Liquid and Solid Impact: ICEAW I, Wear, Vol 233–235, 1999 26. R.H. Barkalow and F.S. Pettit, Corrosion/ Erosion of Materials in Coal Combustion Gas Turbines, Proc. Conf. Corrosion/Erosion of Cal Conversion System Materials (Berkeley, CA), 24–26 Jan 1979, National Association of Corrosion Engineers, 1979, p 139–173 27. D.M. Rishel, F.S. Petit, and N. Birks, Some Principal Mechanisms in the Simultaneous Erosion and Corrosion Attack of Metals at High Temperature, paper 16, Proc. Conf. Corrosion-Erosion-Wear of Materials at Elevated Temperatures (Berkeley, CA), 31 Jan–2 Feb 1990 A.V. Levy, Ed., NACE/ EPRI/LBL/DOE-FE 28. D.J. Stephenson and J.R. Nicholls, Modelling the Influence of Surface Oxidation on High Temperature Erosion, Wear, Vol 186–187, 1995, p 284–290 29. L.K. Ives, Erosion of 310 Stainless Steel at 975 °C in Combustion Gas Atmospheres, Trans. ASME, April 1977, p 126–132 30. M.M. Stack and L. Bray, Interpretation of Wastage Mechanisms of Materials Exposed to Elevated Temperature Erosion-Corrosion Using Erosion-Corrosion Maps and Computer Graphics, Wear, Vol 186–187, 1995, p 273–283 31. M.M. Stack, Looking Beyond the Millennium: Critical Issues in the Evaluation of Materials Performance for Resistance to Erosive Wear in Corrosive Conditions, Wear, Vol 233–235, 1999, p 484– 496 32. J. Dobromirski and I.O. Smith, Metallographic Aspects of Surface Damage, Surface Temperature, and Crack Initiation in Fretting Fatigue, Wear, Vol 117, 1987, p 347–357

33. M.M. Hamdy and R.B. Waterhouse, The Fretting-Fatigue Behavior of a Nickel- Based Alloy (Inconel 718) at Elevated Temperatures, Proc Int. Conf Wear of Materials 1979, American Society of Mechanical Engineers, 1979, p 351–355 34. T.F.J. Quinn, J.L. Sullivan, and D.M. Rowson, Origins and Development of Oxidational Wear at Low Ambient Temperatures, Wear, Vol 94, 1984, p 175–191 35. G. Straffelini, D. Trabucco, and A. Molinari, Oxidative Wear of Heat-Treated Steels, Wear, Vol 250, 2001, p 485–491 36. P.J. Blau, Ed., Friction, Lubrication, and Wear Technology, Vol 18, ASM Handbook, ASM International, 1992 37. “Standard Test Method for Conducting Erosion Tests by Solid Particle Impingement Using Gas Jets,” G 76-02, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM, 2002 38. “Standard Guide for Determining Synergism Between Wear and Corrosion,” G 119- 93, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, ASTM, 1998 39. A. Levy, J. Yan, and J. Patterson, Elevated Temperature Erosion of Steels, Wear, Vol 108, 1986, p 43– 60 40. V. Shanov and W. Tabakoff, Erosion Resistance of Coatings for Metal Protection at Elevated Temperatures, J. Surf. Coat. Technol., Vol 86–87, 1997, p 86–93 41. J. Stringer, Practical Experience with Wastage at Elevated Temperatures in Coal Combustion Systems, Wear, Vol 186–187, 1995, p 11–27


Selected References • • • • • •

R.L. Deuis, C. Subramanian, and J. M. Yellup, Dry Sliding Wear of Aluminum Composites—A Review, Compos. Sci. Technol., Vol 57, 1997, p 415–435 S. Hogmark, Å. Hammarsten, and S. Söderberg, On the Combined Effects of Corrosion and Erosion, Proc. Sixth Int. Conf. on Erosion by Liquid and Solid Impact, 1983, p 37-1 to 37-8 S.C. Lim and M.F. Ashby, Wear-Mechanism Maps, Acta Metall., Vol 35, 1987, p 1–24 F.H. Stott, High-Temperature Sliding Wear of Metals, Tribol. Int., Vol 35, 2002, p 489–495 F.H. Stott, The Role of Oxidation in the Wear of Alloys, Tribol. Int., Vol 31, 1998, p 61–71 G. Sundararajan and M. Roy, Solid Particle Erosion Behavior at Room and Elevated Temperatures, Tribol. Int., Vol 30, 1997, p 339–359

B. Craig, Introduction to Environmentally Induced Cracking, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 345

Introduction to Environmentally Induced Cracking Bruce Craig, MetCorr

Introduction THIS SUBSECTION was developed to introduce the fundamental aspects of environmentally induced cracking. It provides theoretical basis for further discussions on the evaluation, testing, and methods of protection against environmentally induced cracking found in the rest of this Volume. In the time between the publication of the first ASM Handbook on corrosion in 1987 and this publication, considerable work has been done in the area of environmentally induced cracking. However, little gain has been made on the theory of the various cracking mechanisms, and certainly the same is true for progress in modeling or quantitative methods to predict cracking induced by the environment. Thus, the information presented in the previous edition remains fresh today and provides a good starting point for a basic understanding of environmentally induced cracking. Where appropriate, the authors and reviewers of this subsection have updated or added information that brings this subsection up to the current state of the art. The efforts of the original authors are acknowledged at the end of each article.

B. Craig, Introduction to Environmentally Induced Cracking, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 345 Introduction to Environmentally Induced Cracking Bruce Craig, MetCorr

Mechanisms That Induce Cracking This subsection addresses the mechanisms of corrosion that produce cracking of metals and intermetallic compounds as a result of exposure to their environment. This cracking may take the form of relatively slow, stable crack extension with a predictable growth rate or, as is often the case, unpredictable catastrophic fracture. The subsection is divided into four articles (after this brief introduction). Each addresses a specific type of cracking or embrittlement phenomena: stress-corrosion cracking (SCC), hydrogen damage (frequently referred to as hydrogen embrittlement), liquid-metal embrittlement, and solid metal induced embrittlement. In general, these different phenomena show many similarities, and it would at first seem appropriate to propose an allencompassing mechanism to account for these behaviors. For example, all of these phenomena generally are dependent on yield strength and applied stress. As both of these factors increase, resistance to SCC, hydrogen damage, liquid-metal embrittlement, and solid metal induced embrittlement decreases. However, as presented in the discussion of each of these phenomena, many differences between the various forms of environmentally induced cracking are encountered, and in fact, substantial differences are observed for behavior of metals and alloys within a specific form of cracking. At this time, the understanding of each of these mechanisms of cracking is largely phenomenological. No satisfactory theory exists for any of these mechanisms that totally explains all behavior observed either under laboratory or field conditions. Although many theories that are specific to the behavior of certain alloy systems or environments, exist, none is universal enough to explain, for example, the diverse behavior of hydrogen

damage for systems that develop hydrides versus alloys of iron that do not form hydrides. There are many of these contradictory factors that impede the development of an allencompassing theory. Additionally, there has been a continuing controversy since the 1960s concerning the actual micromechanistic causes of SCC, which some investigators consider to be related to hydrogen damage and not strictly an activepath corrosion phenomenon. Although certain convincing data exist for a role of hydrogen in SCC of certain alloys, sufficient data are not available to generalize this concept. As the understanding of these forms of cracking increases, the complexity of each also increases. For example, liquid-metal embrittlement has long been a recognized form of cracking induced by liquid metals. However, it has been shown that certain metals can become embrittlers below their melting point and that embrittlement by these metals becomes progressively worse as the temperature approaches the melting point. This behavior, although similar to liquid-metal embrittlement, is sufficiently different to warrant reclassifying it as solid metal induced embrittlement. Many of the environmentally induced fractures that were previously termed liquidmetal embrittlement are now considered solid metal induced embrittlement. Because satisfactory mechanistic models have not been developed for any of these forms of environmental cracking, the prediction of environmentally assisted cracking is essentially nonexistent. However, the need for prediction of these types of failures is most important, because observable and measurable corrosion usually does not occur before or during crack initiation or propagation. When corrosion does occur, it is highly localized such as pitting or crevice attack and may be difficult to detect. The field of environmentally assisted cracking is one of the most active in corrosion research, and it is continually changing. Therefore, the reader is encouraged to consult the most current literature for more detail on specific theories and alloy behavior. Toward this end, Selected References are given at the end of each article. Additional information also can be found in the articles “Evaluating Stress-Corrosion Cracking” and “Evaluating Hydrogen Embrittlement” in this Volume.

R.H. Jones, Stress-Corrosion Cracking, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 346–366

Stress-Corrosion Cracking R.H. Jones, Battelle Pacific Northwest National Laboratory

Introduction STRESS-CORROSION CRACKING (SCC) describes service failures in engineering materials that occur by slow environmentally induced crack propagation. The observed crack propagation is the result of the combined and synergistic interaction of mechanical stress and corrosion reactions. This is a simple definition of a complex subject, and, like most simplifications, it fails to identify the boundaries of the subject. As a result, before this problem can be discussed in detail, one must clearly define the type of loading involved, the types of materials involved, the types of environments that cause this type of crack propagation, and the nature of the interactions that result in this phenomenon. The term stress-corrosion cracking is used frequently to describe any type of environmentally induced or assisted crack propagation. However, in this discussion, the focus is on the normal usage of the term as defined subsequently. One frequent misconception is that SCC is the result of stress concentration at corrosion-generated surface flaws (as quantified by the stress- intensity factor, K); when a critical value of stress concentration, Kcrit, is reached, mechanical fracture results. Although stress concentration does occur at such flaws, these defects do not exceed the critical value required to cause mechanical fracture of the material in an inert environment (KSCC < Kcrit). Precorrosion followed by loading in an inert environment will not result in any observable crack propagation, while simultaneous environmental exposure and application of stress will cause time-dependent subcritical crack propagation. The term synergy was used to describe this process because the combined simultaneous interaction of mechanical and chemical forces results in crack propagation where neither factor

acting independently or alternately would result in the same effect. The exact nature of this interaction is the subject of numerous scientific investigations and is covered in the section “Crack Propagation Mechanisms” in this article. The stresses required to cause SCC are small, usually below the macroscopic yield stress, and are tensile in nature. The stresses can be externally applied, but residual stresses often cause SCC failures. However, compressive residual stresses can be used to prevent this phenomenon. Static loading usually is considered to be responsible for SCC, while environmentally induced crack propagation due to cyclic loading is defined as corrosion fatigue (see the section “Corrosion Fatigue” in the article “Forms of Mechanically Assisted Degradation” in this Volume). The boundary between these two classes of phenomena is vague, and corrosion fatigue often is considered to be a subset of SCC. However, because the environments that cause corrosion fatigue and SCC are not always the same, these two should be considered separate phenomena. The term stress-corrosion cracking usually is used to describe failures in metallic alloys. However, other classes of materials also exhibit delayed failure by environmentally induced crack propagation. Ceramics exhibit environmentally induced crack propagation (Ref 1), and polymeric materials frequently exhibit craze cracking as a result of the interaction of applied stress and environmental reactions (Ref 2, 3, 4, 5). Until recently, it was thought that pure metals were immune to SCC; however, it has been shown that pure metals are susceptible to SCC (Ref 6, 7). This discussion focuses on the SCC of metals and their alloys and is not concerned with stress- environment interactions in ceramics and polymers. Environments that cause SCC are usually aqueous and can be condensed layers of moisture or bulk solutions. Typically, SCC of an alloy is the result of the presence of a specific chemical species in the environment. Thus, the SCC of copper alloys, traditionally referred to as season cracking, is virtually always due to the presence of ammonia in the environment, and chloride ions cause or exacerbate cracking in stainless steels and aluminum alloys. Also, an environment that causes SCC in one alloy may not cause SCC in another alloy. Changing the temperature, the degree of aeration, and/or the concentration of ionic species may change an innocuous environment into one that causes SCC failure. Also, an alloy may be immune in one heat treatment and susceptible in another. As a result, the list of all possible alloy-environment combinations that cause SCC is expanding continually, and the possibilities are virtually infinite. A partial listing of some of the more commonly observed alloy-environment combinations that result in SCC is given in Table 1. Table 1 Alloy-environment systems exhibiting SCC Alloy Environment Hot nitrate, hydroxide, and carbonate/bicarbonate solutions Carbon steel Aqueous electrolytes, particularly when containing H2S High-strength steels Austenitic stainless steels Hot, concentrated chloride solutions; chloride-contaminated steam High-purity steam High-nickel alloys Ammoniacal solutions a-brass Aqueous Cl-, Br-, and I- solutions Aluminum alloys Aqueous Cl-, Br-, and I- solutions; organic liquids; N2O4 Titanium alloys Aqueous Cl- solutions Magnesium alloys Aqueous Cl- solutions; organic liquids; I2 at 350 °C (660 °F) Zirconium alloys In general, SCC is observed in alloy-environment combinations that result in the formation of a film on the metal surface. These films may be passivating layers, tarnish films, or dealloyed layers. In many cases, these films reduce the rate of general or uniform corrosion, making the alloy desirable for resistance to uniform corrosion in the environment. As a result, SCC is of greatest concern in the corrosion-resistant alloys exposed to aggressive aqueous environments. The exact role of the film in the SCC process is the subject of current research and is covered in the discussion “Crack Propagation Mechanisms” in this article. Table 2 lists some alloy-environment combinations and the films that may form at the crack tip.

Table 2 Alloy-environment combinations and the resulting films that form at the crack tip Metal or alloy Environment α-brass, copper-aluminum Ammonia FeCl3 Gold-copper Acid sulfate Chloride Iron-chromium nickel Hydroxide High-temperature water Nitrite α-brass Nitrite Copper Ammonia (cupric) High-temperature water Ferritic steel Phosphate Anhydrated ammonia CO/CO2/H2O CS2/H2O Chloride Titanium alloys Various media Aluminum alloys, steels

Initiating layer Dealloyed layer (Cu) Dealloyed layer (Au) Dealloyed layer (Au) Dealloyed layer (Ni) Dealloyed layer or oxide Dealloyed layer or oxide Oxide Oxide Porous dissolution zone Oxide Oxide (?) Nitride Carbide Carbide Hydride Near-surface hydrogen

References cited in this section 1. T.A. Michalske and S.W. Freiman, Nature, Vol 295, 11 Feb 1982, p 511 2. J.F. Fellers and B.F. Kee, J. Appl. Polymer Sci., Vol 18, 1974, p 2355 3. E.H. Andrews, Developments in Polymer Science, Applied Science, 1979 4. H.T. Sumison and D.P. Williams, in Fatigue of Composite Materials, STP 569, ASTM, 1975, p 226 5. C. Zeben, in Analysis of the Test Methods for High Modulus Fibers and Composites, STP 521, ASTM, 1973, p 65 6. K. Sieradski and R.C. Newman, Philos. Mag. A, Vol 5 (No. 1), 1985, p 95 7. T. Cassaigne, E.N. Pugh, and J. Kruger, National Bureau of Standards and Johns Hopkins University, unpublished research, 1987

R.H. Jones, Stress-Corrosion Cracking, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 346–366 Stress-Corrosion Cracking R.H. Jones, Battelle Pacific Northwest National Laboratory

The Phenomenon of SCC

Stress-corrosion cracking is a delayed failure process. That is, cracks initiate and propagate at a slow rate from 10-9 to 10-6 m/s (4 × 10-7 to 4 × 10-4 in./s), until the stresses in the remaining ligament of metal exceed the fracture strength. The sequence of events involved in the SCC process usually is divided into three stages: • • •

Crack initiation and stage 1 propagation Stage 2 or steady-state crack propagation Stage 3 crack propagation or final failure

Distinction among these stages is difficult because the transition occurs in a continuous manner and the division is therefore arbitrary. To enhance the understanding of this process, these steps are discussed in the context of typical experimental techniques and results. Stress-corrosion cracking experiments can be classified into three different categories: • • •

Tests on statically loaded smooth samples Tests on statically loaded precracked samples Tests using slowly straining samples

More detailed information on these tests can be found in the article “Evaluating Stress-Corrosion Cracking” in this Volume. Tests on statically loaded smooth samples usually are conducted at various fixed stress levels, and the time to failure of the sample in the environment is measured. Figure 1 illustrates the typical results obtained from this type of test. In Fig. 1, the logarithm of the measured time to failure, tf, is plotted against the applied stress, σapplied, and the time to failure can be seen to increase rapidly with decreasing stress; a threshold stress, σth, is determined where the time to failure approaches infinity. The total time to failure at a given stress consists of the time required for the formation of a crack (the incubation or initiation time, tin, and the time of crack propagation, tcp. These experiments can be used to determine the maximum stress that can be applied to avoid SCC failure, to determine an inspection interval to confirm the absence of SCC crack propagation, or to evaluate the influence of metallurgical and environmental changes on SCC. However, the actual time for the formation or initiation of cracks is strongly dependent on a wide variety of parameters, such as surface finish and prior history. If a cracklike flaw or a crevice is present in the material, then the time to initiate a crack may be reduced dramatically (Ref 8).

Fig. 1 Schematic of a typical time to failure as a function of initially applied stress for smooth sample stress-corrosion cracking tests

Tests on statically loaded precracked samples usually are conducted with either a constant applied load or with a fixed crack opening displacement, and the actual rate or velocity of crack propagation, da/dt, is measured (Ref 9). The magnitude of the stress distribution at the crack tip (the mechanical driving force for crack propagation) is quantified by the stress intensity factor, K, for the specific crack and loading geometry. As a result, the crack propagation rate, da/dt, is plotted versus K as illustrated in Fig. 2. These tests can be configured such that K increases with crack length (constant applied load), decreases with increasing crack length (constant crack mouth opening displacement), or is approximately constant as the crack length changes (special tapered samples). Each type of test has its advantages and disadvantages. However, in service, most SCC failures occur under constant-load conditions such that the stress intensity increases as the crack propagates. As a result, it is usually assumed in SCC discussions that the stress intensity is increasing with increasing crack length.

Fig. 2 Schematic of typical crack propagation rate as a function of crack tip stress-intensity behavior illustrating the regions of stages 1, 2, and 3 crack propagation, as well as identifying the plateau velocity and the threshold stress intensity Typically, three regions of crack propagation rate versus stress-intensity level are found during crack propagation experiments. These regions are identified according to increasing stress-intensity factor as stage 1, 2, or 3 crack propagation (Fig. 2). No crack propagation is observed below some threshold stress-intensity level, KISCC. This threshold stress level is determined not only by the alloy but also by the environment and metallurgical condition of the alloy, and presumably, this level corresponds to the minimum required stress level for synergistic interaction with the environment. At low stress-intensity levels (stage 1), the crack propagation rate increases rapidly with the stress-intensity factor. At intermediate stress-intensity levels (stage 2), the crack propagation rate approaches some constant velocity that is virtually independent of the mechanical driving force. This plateau velocity, Vplateau, is characteristic of the alloy-environment combination and is the result of rate-limiting environmental processes such as mass transport of environmental species up the crack to the crack tip. In stage 3, the rate of crack propagation exceeds the plateau velocity as the stress-intensity level approaches the critical stress-intensity level for mechanical fracture in an inert environment, Klc (Ref 10). Slow Strain Rate Testing. Stress-corrosion tests also can be conducted by slowly increasing the load or strain on either precracked or smooth samples. Usually, a tensile machine pulls a smooth sample that is exposed to the corrosive environment at a low cross head speed, 10-5 to 10-9 m/s. The strain to failure in the corrosive environment and the strain to failure in an inert environment can then be plotted against the strain rate, as shown in Fig. 3(a), or the ratio of these measurements can be plotted as shown in Fig. 3(b) for ductility. The ratio of other tensile property measurements, such as reduction in area and ultimate tensile strength, could also be plotted. These tests are known as constant extension rates tests (CERT), slow strain rate tests, or straining electrode tests. These tests are excellent for comparing the relative susceptibility of alloys to cracking in an

environment or for studying the influence of metallurgical variables on the susceptibility of an alloy. However, the application of this data to the prediction of actual in-service crack propagation rates is difficult and unreliable.

Fig. 3 Strain to failure plots resulting from slow strain rate testing. (a) Ductility of two alloys is measured by elongation, reduction in area, or fracture energy in the aggressive environment and an inert reference environment. (b) Schematic of typical ductility ratio of the effects of the aggressive over the inert reference environment versus strain rate behavior of two different types of alloys

References cited in this section 8. C.M. Chen, M.H. Froning, and E.D. Verink, in Stress Corrosion—New Approaches, STP 610, ASTM, 1976, p 289 9. C.J. Beevers, Ed., The Measurement of Crack Length and Shape During Fracture and Fatigue, Chameleon Press, 1980 10. E.N. Pugh, U. Bertocci, and R.E. Ricker, National Bureau of Standards, unpublished research, 1987


The Phenomenon of SCC Stress-corrosion cracking is a delayed failure process. That is, cracks initiate and propagate at a slow rate from 10-9 to 10-6 m/s (4 × 10-7 to 4 × 10-4 in./s), until the stresses in the remaining ligament of metal exceed the fracture strength. The sequence of events involved in the SCC process usually is divided into three stages: • • •

Crack initiation and stage 1 propagation Stage 2 or steady-state crack propagation Stage 3 crack propagation or final failure

Distinction among these stages is difficult because the transition occurs in a continuous manner and the division is therefore arbitrary. To enhance the understanding of this process, these steps are discussed in the context of typical experimental techniques and results. Stress-corrosion cracking experiments can be classified into three different categories: • • •

Tests on statically loaded smooth samples Tests on statically loaded precracked samples Tests using slowly straining samples

More detailed information on these tests can be found in the article “Evaluating Stress-Corrosion Cracking” in this Volume. Tests on statically loaded smooth samples usually are conducted at various fixed stress levels, and the time to failure of the sample in the environment is measured. Figure 1 illustrates the typical results obtained from this type of test. In Fig. 1, the logarithm of the measured time to failure, tf, is plotted against the applied stress, σapplied, and the time to failure can be seen to increase rapidly with decreasing stress; a threshold stress, σth, is determined where the time to failure approaches infinity. The total time to failure at a given stress consists of the time required for the formation of a crack (the incubation or initiation time, tin, and the time of crack propagation, tcp. These experiments can be used to determine the maximum stress that can be applied to avoid SCC failure, to determine an inspection interval to confirm the absence of SCC crack propagation, or to evaluate the influence of metallurgical and environmental changes on SCC. However, the actual time for the formation or initiation of cracks is strongly dependent on a wide variety of parameters, such as surface finish and prior history. If a cracklike flaw or a crevice is present in the material, then the time to initiate a crack may be reduced dramatically (Ref 8).

Fig. 1 Schematic of a typical time to failure as a function of initially applied stress for smooth sample stress-corrosion cracking tests Tests on statically loaded precracked samples usually are conducted with either a constant applied load or with a fixed crack opening displacement, and the actual rate or velocity of crack propagation, da/dt, is measured (Ref 9). The magnitude of the stress distribution at the crack tip (the mechanical driving force for crack propagation) is quantified by the stress intensity factor, K, for the specific crack and loading geometry. As a result, the crack propagation rate, da/dt, is plotted versus K as illustrated in Fig. 2. These tests can be configured such that K increases with crack length (constant applied load), decreases with increasing crack length (constant crack mouth opening displacement), or is approximately constant as the crack length changes (special tapered samples). Each type of test has its advantages and disadvantages. However, in service, most SCC failures occur

under constant-load conditions such that the stress intensity increases as the crack propagates. As a result, it is usually assumed in SCC discussions that the stress intensity is increasing with increasing crack length.

Fig. 2 Schematic of typical crack propagation rate as a function of crack tip stress-intensity behavior illustrating the regions of stages 1, 2, and 3 crack propagation, as well as identifying the plateau velocity and the threshold stress intensity Typically, three regions of crack propagation rate versus stress-intensity level are found during crack propagation experiments. These regions are identified according to increasing stress-intensity factor as stage 1, 2, or 3 crack propagation (Fig. 2). No crack propagation is observed below some threshold stress-intensity level, KISCC. This threshold stress level is determined not only by the alloy but also by the environment and metallurgical condition of the alloy, and presumably, this level corresponds to the minimum required stress level for synergistic interaction with the environment. At low stress-intensity levels (stage 1), the crack propagation rate increases rapidly with the stress-intensity factor. At intermediate stress-intensity levels (stage 2), the crack propagation rate approaches some constant velocity that is virtually independent of the mechanical driving force. This plateau velocity, Vplateau, is characteristic of the alloy-environment combination and is the result of rate-limiting environmental processes such as mass transport of environmental species up the crack to the crack tip. In stage 3, the rate of crack propagation exceeds the plateau velocity as the stress-intensity level approaches the critical stress-intensity level for mechanical fracture in an inert environment, Klc (Ref 10). Slow Strain Rate Testing. Stress-corrosion tests also can be conducted by slowly increasing the load or strain on either precracked or smooth samples. Usually, a tensile machine pulls a smooth sample that is exposed to the corrosive environment at a low cross head speed, 10-5 to 10-9 m/s. The strain to failure in the corrosive environment and the strain to failure in an inert environment can then be plotted against the strain rate, as shown in Fig. 3(a), or the ratio of these measurements can be plotted as shown in Fig. 3(b) for ductility. The ratio of other tensile property measurements, such as reduction in area and ultimate tensile strength, could also be plotted. These tests are known as constant extension rates tests (CERT), slow strain rate tests, or straining electrode tests. These tests are excellent for comparing the relative susceptibility of alloys to cracking in an environment or for studying the influence of metallurgical variables on the susceptibility of an alloy. However, the application of this data to the prediction of actual in-service crack propagation rates is difficult and unreliable.

Fig. 3 Strain to failure plots resulting from slow strain rate testing. (a) Ductility of two alloys is measured by elongation, reduction in area, or fracture energy in the aggressive environment and an inert reference environment. (b) Schematic of typical ductility ratio of the effects of the aggressive over the inert reference environment versus strain rate behavior of two different types of alloys

References cited in this section 8. C.M. Chen, M.H. Froning, and E.D. Verink, in Stress Corrosion—New Approaches, STP 610, ASTM, 1976, p 289 9. C.J. Beevers, Ed., The Measurement of Crack Length and Shape During Fracture and Fatigue, Chameleon Press, 1980 10. E.N. Pugh, U. Bertocci, and R.E. Ricker, National Bureau of Standards, unpublished research, 1987 R.H. Jones, Stress-Corrosion Cracking, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 346–366 Stress-Corrosion Cracking R.H. Jones, Battelle Pacific Northwest National Laboratory

Controlling Parameters The mechanisms that have been proposed for SCC require that certain processes or events occur in sequence for sustained crack propagation to be possible. These requirements explain the plateau region in which the rate of crack propagation is independent of the applied mechanical stress. That is, a sequence of chemical reactions and processes is required, and the rate-limiting step in this sequence of events determines the limiting rate or plateau velocity of crack propagation (until mechanical overload fracture starts contributing to the fracture process in stage 3). Figure 4 illustrates a crack tip in which crack propagation results from reactions in metal ahead of the propagating crack. Close examination of Fig. 4 reveals that potential rate-determining steps include: • • • •

Mass transport along the crack to the crack tip Reactions in the solution near the crack Surface adsorption at or near the crack tip Surface diffusion

• • • • •

Surface reactions Absorption into the bulk Bulk diffusion to the plastic zone ahead of the advancing crack Chemical reactions in the bulk The rate of interatomic bond rupture

Fig. 4 Schematic of crack tip processes that may be the rate-determining step in environmentally assisted crack propagation Changes in the environment that modify the rate-determining step will have a dramatic influence on the rate of crack propagation, while alterations to factors not involved in the rate-determining step or steps will have little influence, if any. However, significantly retarding the rate of any one of the required steps in the sequence could make that step the rate-determining step. In aqueous solutions, the rate of adsorption and surface reactions is usually very fast compared with the rate of mass transport along the crack to the crack tip. As a result, bulk transport into this region or reactions in this region frequently are believed to be responsible for determining the steady-state crack propagation rate or plateau velocity. In gaseous environments, surface reactions, surface diffusion, and adsorption may be rate limiting, as well as the rate of bulk transport to the crack tip (Ref 11, 12). Several different environmental parameters are known to influence the rate of crack growth in aqueous solutions. These parameters include, but are not limited to: • • • • • • • •

Temperature Pressure Solute species Solute concentration and activity pH Electrochemical potential Solution viscosity Stirring or mixing

Altering any of these parameters may alter the rate of the rate-controlling steps, either accelerating or reducing the rate of crack propagation. Also, it may be possible to arrest or stimulate crack propagation by altering the rate of an environmental reaction. It is well known and generally accepted that the environment at occluded sites, such as a crack tip, can differ significantly from the bulk solution. If an alteration to the bulk environment allows the formation of a critical SCC environment at crack nuclei, then crack propagation will result. If the bulk environment cannot maintain this local crack tip environment, then crack propagation will stop. As a result, slight changes to the environment may have a dramatic influence on crack propagation, while dramatic changes may have only a slight influence.

In addition to these environmental parameters, stress-corrosion crack propagation rates are influenced by: • • • • • •

The magnitude of the applied stress or the stress-intensity factor The stress state, which includes plane stress and plane strain The loading mode at the crack tip (tension or torsion, for example) Alloy composition, which includes nominal composition, exact composition (all constituents), and impurity or tramp element composition Metallurgical condition, which includes strength level, second phases present in the matrix and at the grain boundaries, composition of phases, grain size, grain-boundary segregation, and residual stresses Crack geometry, which includes length, width, and aspect ratio; and crack opening and crack tip closure

References cited in this section 11. D.P. Williams and H.G. Nelson, Metall. Trans., Vol 1 (No. 1), 1970, p 63 12. R.E. Ricker and D.J. Duquette, The Role of Environment on Time Dependent Crack Growth, Micro and Macro Mechanics of Crack Growth, K. Sadananda, B.B. Rath, and D.J. Michel, Ed., American Institute of Mining, Metallurgical, and Petroleum Engineers, 1982, p 29


Important Fracture Features Stress-corrosion cracks can initiate and propagate with little outside evidence of corrosion and no warning as catastrophic failure approaches. The cracks initiate frequently at surface flaws that are either preexisting or formed during service by corrosion, wear, or other processes. The cracks then grow with little macroscopic evidence of mechanical deformation in metals and alloys that are normally quite ductile. Crack propagation can be either intergranular or transgranular; sometimes, both types are observed on the same fracture surface. Crack openings and the deformation associated with crack propagation may be so small that the cracks virtually are invisible except in special nondestructive examinations. As the stress intensity increases, the plastic deformation associated with crack propagation increases, and the crack opening increases. When the final fracture region is approached, plastic deformation can be appreciable because corrosion-resistant alloys are frequently quite ductile. The features of stress-corrosion fracture surfaces are covered in greater detail in the article “Modes of Fracture” and the “Atlas of Fractographs” in Fractography, Volume 12 of ASM Handbook, formerly 9th Edition Metals Handbook.


Phenomenology of Crack Initiation Processes Crack Initiation at Surface Discontinuities. Stress-corrosion cracking frequently initiates at preexisting or corrosion-induced surface features. These features may include grooves, laps, or burrs resulting from fabrication processes. Examples of such features are shown in Fig. 5; these features were produced during grinding in the preparation of a joint for welding. The feature shown in Fig. 5(a) is a lap, which subsequently recrystallized during welding and could now act as a crevice at which deleterious cations concentrate. The highly sensitized recrystallized material could also more readily become the site of crack initiation by intergranular corrosion. A cold-worked layer and surface burrs, shown in Fig. 5(b), can also assist crack initiation.

Fig. 5 Optical micrographs showing defects on the inner surface of type 304 stainless steel pipe near weld root (a) and near through crack (b). Both 670× Crack Initiation at Corrosion Pits. Stress- corrosion cracks can also initiate at pits that form during exposure to the service environment (Fig. 6) or by prior cleaning operations, such as pickling of type 304 stainless steel before fabrication. Pits can form at inclusions that intersect the free surface or by a breakdown in the protective film. In electrochemical terms, pits form when the potential exceeds the pitting potential. It has been shown that the SCC potential and pitting potential were identical for steel in nitrite solutions (Ref 13).

Fig. 6 Stress-corrosion crack initiating from a corrosion pit in a quenched-and-tempered high-strength turbine disk steel (3.39Ni-1.56Cr-0.63Mo-0.11V) test coupon exposed to oxygenated, demineralized water for 800 h under a bending stress of 90% of the yield stress. (a) 185×. (b) 248×. Courtesy of S.J. Lennon, ESCOM, and F.P.A. Robinson, University of the Witwatersrand The transition between pitting and cracking is dependent on the same parameters that control SCC—that is, the electrochemistry at the base of the pit, pit geometry, chemistry of the material, and stress or strain rate at the base of the pit. A detailed description of the relationship between these parameters and crack initiation has not been developed because of the difficulty in measuring crack initiation. Methods for measuring short surface cracks are under development but are limited to detecting cracks that are beyond the initiation stage. The pit geometry is important in determining the stress and strain rate at the base of the pit. Generally, the aspect ratio between the penetration and the lateral corrosion of a pit must be greater than approximately 10 before a pit acts as a crack initiation site. A penetration to lateral corrosion ratio of 1 corresponds to uniform corrosion, and a ratio of about 1000 is generally observed for a growing stress- corrosion crack. As in the case of a growing crack, the pit walls must exhibit some passive film-forming capability in order for the corrosion ratio to exceed 1. A change in the corrosive environment and potential within a pit may also be necessary for the pit to act as a crack initiator. Pits can act as occluded cells similar to cracks and crevices, although in general their volume is not as restricted. There are a number of examples in which stress-corrosion cracks initiated at the base of a pit by intergranular corrosion. In these circumstances, the grain-boundary chemistry and the pit chemistry were such that intergranular corrosion was favored. Crack propagation was also by intergranular SCC in these cases. Although the local stresses and strain rates at the base of the pit play a role in SCC initiation, there are examples of preexisting pits that did not initiate stress-corrosion cracks. This observation has led to the conclusion that the electrochemistry of the pit is more important than the local stress or strain rate (Ref 13). A preexisting pit may not develop the same local electrochemistry as one grown during service because the development of a concentration cell depends on the presence of an actively corroding tip that establishes the anion and cation current flows. Similarly, an inability to reinitiate crack growth in samples in which active growth was occurring before the samples were removed from solution, rinsed, dried, and reinserted into solution also suggests that the local chemistry is very important. Crack Initiation by Intergranular Corrosion or Slip Dissolution. Stress-corrosion crack initiation can also occur in the absence of pitting by intergranular corrosion or slip-dissolution processes. Intergranular corrosioninitiated SCC requires that the local grain-boundary chemistry differ from the bulk chemistry. This condition occurs in sensitized austenitic stainless steels or with the segregation of impurities such as phosphorus, sulfur, or silicon in a variety of materials. Slip-dissolution-initiated SCC results from local corrosion at emerging slip planes and occurs primarily in low stacking fault materials. The processes of crack initiation and propagation by the slip-dissolution process are in fact very similar.

Reference cited in this section 13. R.M. Parkins, Mater. Sci. Technol., Vol 1, 1985, p 480


Crack Initiation Mechanisms Although the features causing SCC initiation, such as pits, fabrication defects, and intergranular corrosion, are observed and identified easily, there are few well-developed models of SCC initiation. This lack of models for crack initiation mechanisms is the result of several complicating factors. For example, the initiation of a crack is difficult to measure experimentally, even though it is not difficult to detect the location from which a growing crack has emanated. Furthermore, crack initiation has not been defined precisely, and it is difficult to determine at what point a pit is actually a crack and when intergranular corrosion becomes intergranular SCC. Also, the fracture mechanics concept of design assumes preexisting flaws in materials, although these may not be surface flaws that can become stress-corrosion cracks. It has been demonstrated that the corrosion fatigue threshold of 12% Cr steels and 2.0% NiCrMoV steels could be related to the minimum depth of surface pits, as shown in Fig. 7. Using a linear-elastic fracture mechanics approach and relating a pit to a half-elliptical surface crack, one researcher has shown that the critical pit dimension could be expressed by the following relationship (Ref 14):

Fig. 7 Relationship between stress amplitude and minimum depth of surface defects for 12% Cr steel and a 2.0% NiCrMoV steel. Source: Ref 14

(Eq 1) where ΔKth is the corrosion fatigue threshold, F is a constant, and Δσ0 is the alternating surface stress. A pit could be represented by a half-elliptical surface crack because it had intergranular corrosion at the base that caused the pit to have cracklike characteristics. A model for stress-corrosion crack growth was developed from pits in brass tested in a non- tarnishing ammoniacal solution (Ref 15). The researchers evaluated the corrosion and stress aspects of a pit to develop a model for crack initiation. They assumed that the corrosion conditions in the base of the pit were essentially the same as those on a flat surface and that initiation required that a critical crack tip opening displacement be exceeded. Using linear-elastic fracture mechanics, the researchers developed the following relationship for the time to initiate a crack: (Eq 2) where KISCC is the stress-corrosion threshold for crack growth for brass in ammoniacal solution, σ is the applied stress, σ0 is the stress needed to close the crack, B is a constant, -Vm is the electrochemical potential of the sample, and V0 is the reversible potential. However, the assumption that the pit corrosion conditions are equal to the flat surface conditions limits the applicability of this model, because generally it is accepted that the electrochemistry in a pit differs considerably from that of a flat surface. Also, this model does not describe explicitly the transition from a pit to a crack but treats a pit as a small crack in which the crack tip opening and crack depth are affected by corrosion. The transition between intergranular corrosion and intergranular SCC was evaluated for nickel with segregated phosphorus and sulfur (Ref 16, 17). Because phosphorus and sulfur inhibit the formation of a passive film on nickel, it was expected that intergranular SCC would be similar in these two materials. It was observed, however, that only shallow intergranular corrosion to a depth of about 0.1 mm (0.004 in.) resulted in nickel in which sulfur was segregated to the grain boundary when tested at a strain rate of 10-6 s-1 in 1 N sulfuric acid (H2SO4). On the other hand, rapid intergranular SCC resulted in nickel in which phosphorus was segregated to the grain boundaries that was tested under similar conditions. Using crack tip chemistry modeling, it was shown that intergranular corrosion did not convert to a stress-corrosion crack in Ni+S because the sulfur remained on the crack walls, causing the electrolyte in the solution to become saturated and the crack tip corrosion rate to drop to zero. Starting from a flat surface, the corrosion current at the tip of the intergranular corrosion penetration was about 30 mA/cm2 (195 mA/in.2) and decreased to 0 at a depth of about 0.08 mm (0.003 in.), as shown in Fig. 8. A similar corrosion depth was measured for samples held at 900 mV for extended periods or for straining electrode samples, while for Ni+P a crack growth rate of 10-4 mm/s was measured. These results indicate the importance of electrochemistry in the base of a pit or an intergranular penetration and show that the occurrence of intergranular corrosion does not a priori indicate that intergranular SCC will follow.

Fig. 8 Effect of crack depth on crack tip corrosion rates for nickel with actively corroding crack walls. θ, angle between the crack walls

References cited in this section 14. L. Hagn, Lifetime Prediction for Parts in Corrosion Environments, Corrosion in Power Generating Equipment, Plenum Press, 1983 15. O. Buck and R. Ranjan, Evaluation of a Crack-Tip-Opening Displacement Model Under StressCorrosion Conditions, Modeling Environmental Effects on Crack Growth Processes, R.H. Jones and W.W. Gerberich, Ed., The Metallurgical Society, 1986, p 209 16. M.J. Danielson, C. Oster, and R.H. Jones, Corros. Sci., Vol 32 (No 1), 1991, p 1–21 17. R.H. Jones, M.J. Danielson, and D.R. Baer, J. Mater. Energy Syst., Vol 8, 1986, p 185


Phenomenology of Crack Propagation Processes Crack initiation and propagation are related but different processes; however, by definition, if a crack initiates, it will propagate. As indicated in the section “Crack Initiation Mechanisms” in this article, cracks may initiate at preexisting surface flaws if the necessary electrochemical, mechanical, and metallurgical conditions are met, or corrosion processes may create a surface flaw by pitting or localized corrosion. The conditions that create a pit or localized corrosion at chemical inhomogeneities such as grain boundaries, inclusions, second phases, and interphase boundaries are not necessarily the same as those needed for sustained crack growth. For example, the electrochemical conditions near the surface of a material are similar to the bulk electrolyte conditions; electrolyte conditions in a propagating crack generally differ from the bulk electrolyte. If the conditions at the tip of the growing pit or localized corrosion penetration cannot achieve the proper pH, potential, or chemistry, crack propagation may not proceed. Also, if the pit aspect ratio is not greater than a critical value, then the local stresses and strains resulting from a sustained sharp crack are not attained and crack growth cannot occur. Lastly, for stress-corrosion crack growth resulting from cathodic hydrogen, a short crack may have to initiate by an anodic dissolution process, while propagation depends on hydrogen activity. Although there are conditions under which local corrosion or pitting may not result in crack propagation, more frequently they do lead to cracking. Three conditions cited often as requirements for sustained SCC are a susceptible material, a corrosive environment, and an adequate stress. The list of susceptible material conditions and corrosive environments that are known to cause SCC has been expanding steadily. For example, pure metals are known to crack when impurities are segregated to grain boundaries, nonsensitized stainless steels crack in high-purity water, and ferritic steels crack in environments other than nitrate and caustic solutions. Similarly, the details of the microstructure and microchemistry of a susceptible material, specific ions and chemistry, pH, potential, the corrosion rates of the local crack tip associated with a corrosive environment, and the crack tip stresses, strains, and strain rates associated with cracking are better known today than they were previously. Knowledge and understanding of the factors and mechanisms of stress corrosion are continuing to expand; however, a number of factors are well established. The phenomenological description of stress corrosion based on these known factors is presented in this section.

Environmental Factors Environmental effects on SCC are summarized frequently simply by listing alloy-environment combinations in which SCC has been observed, such as that in Table 1. In recent years, the number of such combinations has increased. Added to this list is the observation of the transgranular SCC of copper, the intergranular SCC of pure metals such as iron and nickel, and the SCC of materials in high-purity water in the absence of specific anions. Stress-corrosion cracking of pure metals was thought to be impossible in early lists of materialenvironment combinations, but it is now realized that specific environment/metal reactions occur that allow crack advance. In the case of intergranular SCC of iron and nickel, this reaction occurs between the environment and impurities segregated to the grain boundaries (a condition that exists in many pure metals and engineering alloys). This increase in known susceptible materials can be attributed to the use of new SCC tests, refined crack propagation monitoring equipment, improved electrochemical control equipment, and, perhaps most importantly, the increased research activity in this field over the last 20 to 30 years. Expectations for material performance have increased also with time as materials are used in more aggressive environments under more demanding loading conditions. Lists such as that given in Table 1 can be useful in materials selection for a design involving corrosive environments because it can lead the materials engineer to seek more specific information on the materials and environments in question. However, such lists can also be misleading because service conditions may differ markedly from those in which the susceptibility listed in Table 1 was determined. Because SCC is dependent on bulk alloy chemistry, microstructure, microchemistry, loading parameters, and specific environmental factors such as oxidizing potential and pH, lists of the type given in Table 1 should be used only for a general overview of SCC. A complete description of SCC must treat both the thermodynamic requirements and kinetic aspects of cracking. A knowledge of the thermodynamic conditions will help determine whether cracking is feasible; kinetic information describes the rate at which cracks propagate. There are thermodynamic requirements for both anodically assisted SCC (where crack propagation is dependent either directly or indirectly on the oxidation of metal atoms from the crack tip and their dissolution in the electrolyte) and for hydrogen-assisted crack growth resulting from the reduction of hydrogen ions at the crack tip. Thermodynamics of SCC. The thermodynamic conditions for anodically assisted SCC are that dissolution or oxidation of the metal and its dissolution in the electrolyte must be thermodynamically possible and that a protective film, such as an oxide or salt, must be thermodynamically stable. The first condition becomes a requirement because, without oxidation, crack advance by dissolution would not result. That a process is controlled by anodic dissolution does not a priori indicate that the total crack extension is the sum of the total number of coulombs of charge exchanged at the crack tip. There are crack advance processes in which the crack advance is controlled by anodic dissolution but in which the total crack length is greater than can be accounted for by the total charge transfer. These processes are covered in the section “Crack Propagation Mechanisms” in this article. However, it is important to note that if the brittle crack advance process is initiated and controlled by anodic dissolution, the crack growth rate will be zero if the anodic current density is zero and will increase with increasing current density. This also holds true for crack advance resulting from brittle crack jump. The thermodynamic requirement of simultaneous film formation and oxidation for stress- corrosion crack growth can be understood from the diagram shown in Fig. 9, in which the ratio of the corrosion currents from the walls relative to the crack tip is the critical parameter. This ratio must be substantially less than 1 for a crack to propagate; otherwise, the crack will blunt, or the crack tip solution will saturate. Crack initiation can also be controlled by this ratio because a pit with a high wall corrosion rate will broaden as fast or faster than it will penetrate, resulting in general corrosion rather than crack growth. It is believed generally that the activity of the crack walls relative to the crack tip is a consequence of greater dynamic strain at the tip than along the walls.

Fig. 9 Schematic of stress-corrosion crack showing important transport and corrosion reactions. Arepresents negatively charged anions migrating to the crack tip, crack solution from the crack walls, and crack tip.

represents metal ions entering the

indicates metal ions entering the crack solution from the

A thermodynamic requirement of simultaneous film formation and oxidation of the underlying material led to the identification of critical potentials for the presence or absence of SCC. An example of these critical potentials is shown in Fig. 10 for a passive film-forming material such as stainless steel. Zones 1 and 2 are those in which transgranular stress-corrosion crack growth is most likely to occur; intergranular stress-corrosion crack growth can occur over a wider range of potentials than these two zones can. Transgranular SCC occurs in zone 1 because the material is in transition from active corrosion to passive film formation such that the simultaneous conditions for film formation on the crack walls and corrosion at the crack tip are met. A similar condition exists in zone 2, with the added factor that these potentials are at or above the pitting potential so that cracks can initiate by pitting.

Fig. 10 Potentiokinetic polarization curve and electrode potential values at which stress-corrosion cracking appears

Intergranular SCC occurs over a wider range of potentials than those shown for zones 1 and 2 because chemical inhomogeneities at the grain boundary produce a different electrochemical response relative to the bulk material. Therefore, passive crack walls and active crack tips can result over the potential range from zone 1 to zone 2. Examples of the critical potentials for SCC are shown in Fig. 11 for several materials. Identification of critical potentials for SCC has led to the use of electrochemical methods for assessing stress-corrosion susceptibility. Zones 1 and 2 are identified by determining the electrochemical potential versus current curves, as shown in Fig. 10 and 11. The shapes of these curves determined at high and low sweep rates are also used to indicate potentials at which the simultaneous conditions of film formation and metal oxidation occur.

Fig. 11 Potentiokinetic polarization curve and electrode potential values at which intergranular and transgranular stress-corrosion cracking appear in a 10% NaOH solution at 288 °C (550 °F). (a) Alloy 600. (b) Alloy 800. (c) AISI type 304 stainless steel Application of Potential-pH Diagrams. Critical potentials for SCC can also be related to potential-pH stability diagrams (Pourbaix diagrams) because these diagrams describe the conditions at which film formation and metal oxidation will occur. An example for carbon steel is given in Fig. 12, in which SCC is associated with potentials and pHs at which phosphate, carbonate, or magnetite films are thermodynamically stable while the species Fe2+ and are metastable. A second example of the potential-pH regimes in which SCC occurs is given in Fig. 13 for a 70Cu-30Zn brass in a variety of solutions.

Fig. 12 Relationship between pH-potential conditions for severe cracking susceptibility of carbon steel in various environments and the stability regions for solid and dissolved species on the Pourbaix diagram. Note that severe susceptibility is encountered where a protective film (phosphate, carbonate, magnetite, etc.) is thermodynamically stable, but if ruptured, a soluble species (Fe2+, ) is metastable.

Fig. 13 Potential-pH diagram showing the domains of failure mode for 70Cu-30Ni brass in various solutions, together with the calculated positions of various boundaries relating to the domains of stability of different chemical species The effect of many environmental parameters, such as pH, oxygen concentration, and temperature, on the thermodynamic conditions for SCC can be related to their effect on the potential-pH diagrams (as shown in Fig. 12 and 13) or on the material potential relative to the various stability regions. Using the potential-pH diagram for iron in water at 25 °C (77 °F) shown in Fig. 14, the effects of changing the pH and oxygen concentration can be illustrated. A decrease in the pH from 9 to 6 at a potential of -0.2 V versus standard hydrogen electrode (SHE) shifts iron from a region of stability to one of active corrosion. Based on the thermodynamic criteria for simultaneous stability and active corrosion, the critical pH would be expected to be 7, with decreasing susceptibility at higher pHs because of increased film stability and at lower pHs because of decreased film stability and the increase of general corrosion. Changes in the oxygen concentration generally alter the electrode potential, with increasing oxygen concentration resulting in more oxidizing conditions. The effects of temperature on the potential-pH diagram must be determined for each temperature of interest as the regions of stability shift with temperature.

Fig. 14 Potential-pH diagram (Pourbaix) for iron in water at 25 °C (77 °F). A decrease in pH from 9 to 6 at potential of -0.2 V, which shifts iron from a region of stability to one of active corrosion, is indicated by the solid bar. For materials in which SCC occurs by a hydrogen-induced subcritical crack growth mechanism, the thermodynamic requirement for crack growth is governed by the hydrogen reduction line shown by dashed line a in Fig. 14. Hydrogen reduction on iron in water at 25 °C (77 °F) occurs at potentials below this line, but not above it. Therefore, the range of potentials at which hydrogen is available to cause crack growth increases and becomes more oxidizing with decreasing pH. Application of potential-pH diagrams for identifying specific conditions at which SCC will occur is limited by a number of factors, such as the availability of these diagrams for complex solutions and for the temperatures of interest and the substantial deviation of the chemistry of a crevice or crack from the bulk solution chemistry. Also, the electrode potential at the crack tip can differ from that of the free surface of the material. These differences arise from the need for diffusion of metal ions away from the actively corroding crack tip, migration of anions into the crack, reactions along the crack walls, convection of the electrolyte in the crack, reactions along the crack walls, and, in some cases, by gas bubble formation that causes potential drops. Efforts to measure the local crack tip chemistry and potentials are restricted by crack sizes, which can be substantially smaller than 1 μm. Therefore, mathematical modeling of transport within cracks and the resulting crack tip chemistry, reactions, and potentials have been pursued actively in recent years. The results of measurements in simulated crevices and modeling have shown clearly that the crack tip chemistry can differ from the bulk conditions, but the specifics of these differences are not well known and depend on the material and environment being considered. For example, based on evaluations by one researcher, the pH within cracks and crevices of structural steels in marine environments ranged from 4 to 11, of stainless steel from 0 to 3, of aluminum from 3 to 4, and of titanium from 1 to 2 (Ref 18). Shifts of up to 300 mV in the crack tip potential in the anodic direction were observed for cathodically polarized steel, while other researchers found only a small shift in the crack tip potential relative to the external potential under steady-state conditions (Ref 16) or a 300 mV shift in the cathodic direction when the crevice walls were active, reducing to a small shift when the walls passivated (Ref 19). Therefore, the critical potentials at which cracking may occur can differ from those given by the bulk equilibrium conditions used to determine the potential-pH diagrams. Although it is possible to determine the equilibrium conditions based on the crack tip conditions, the current knowledge of crack tip conditions is fragmentary at best. Other factors that limit the usefulness of the equilibrium potential-pH diagrams are the lack of available diagrams for specific materials and environments and the variable temperatures, chemistries, and potentials common in many industrial applications. Kinetics of SCC. A knowledge of the thermodynamic conditions at which SCC can occur is insufficient without a corresponding understanding of the kinetics of crack growth because the life of a component may be adequate if the crack growth rate is sufficiently slow, even though SCC is thermodynamically possible. As in the thermodynamic conditions for SCC, environmental parameters such as potential, pH, oxygen concentration, temperature, and crack tip chemistry have a strong effect on the crack growth kinetics. The crack tip reactions and the rate-determining steps controlling the crack growth rate are specific to alloy-environment combinations; although a discussion of each system is beyond the scope of this section, some general observations are made. Also, the crack growth rate depends on the crack advance process even though it is controlled by electrochemical reactions. Detailed descriptions of various crack growth mechanisms are given in the section “Crack Propagation Mechanisms” in this article by categorizing these mechanisms as either anodic dissolution or mechanical fracture models. For the case of a crack growing by anodic dissolution alone, the total crack advance is a function of the total charge transfer at the crack tip, while the crack velocity is a function of the crack tip current density. For a crack growing with mechanical fracture, the total crack advance exceeds the total charge transfer at the crack tip, but the crack velocity may still be controlled by the crack tip current density. A limiting velocity, da/dt, can be described for a crack advancing under pure anodic dissolution by the following Faradaic relationship: (Eq 3)

where ia is the anodic current density of a bare surface, M is the atomic weight, z is the valence, F is Faraday's constant, and ρ is the material density. It has been shown (Ref 20) that this relationship (Eq 3) between the bare surface current density and the crack propagation rate is applicable to a wide variety of materials, as presented in Fig. 15. Equation 2 assumes that the crack tip is maintained in a bare condition, while the crack walls are relatively inactive. A bare crack tip and passive walls can result from the difference between the electrochemical conditions at the tip and other regions in the crack, a difference in the local chemistry of the material at the crack tip that causes the tip to be more active than the crack walls (such as at a sensitized or segregated grain boundary), or a crack tip strain rate that is sufficiently high to prevent the formation of a protective film.

Fig. 15 Relationship between the average crack propagation rate and the oxidation (that is, dissolution and oxide growth) kinetics on a straining surface for several ductile alloy/aqueous environment systems A number of factors can reduce the crack velocity below that given by Eq 3 and in Fig. 15. The most widely examined crack growth retardation process is that resulting from the crack tip being covered by a film for some fraction of time. The process of crack growth in the presence of a film at the crack tip has been described by several mechanisms, such as slip dissolution and passive film rupture, covered in the section “Crack Propagation Mechanisms” in this article. In general, the crack growth rate depends on the rate at which the film

is ruptured and reformed. The amount of corrosion that occurs between these two events has been used to describe the crack growth rate. This time period is determined by the crack tip strain rate, film fracture strain, repassivation rate of the surface, maximum corrosion rate while the tip is bare, and the corrosion rate decay with repassivation. Other factors that can reduce the crack growth rate below that given by Eq 3 are limits in the diffusion rate of species into and out of the crack tip, crack deflection away from the principal stress, and changes in the local material chemistry. Transport of species into and out of cracks is considered a major limitation to attaining the crack growth rate predicted by Eq 3. Factors such as the crack geometry or width, reactions or corrosion rate along the walls of the crack, diffusion rate of anions and cations, and metal salt solubility limits all contribute to transport-limited crack velocities. A lack of clear knowledge about the specific conditions at the tip of a crack has limited the understanding of the role of specific species on crack growth rates. At this time, it is possible only to describe the effect of bulk electrolyte conditions on crack growth rate; the local conditions may vary significantly. However, knowledge of the local crack tip conditions is most important for understanding the mechanisms of cracking, although a knowledge of the external crack conditions is adequate for monitoring and controlling SCC. Using austenitic stainless steel as an example, the effects of electrochemical potential, oxygen content, and temperature are shown schematically in Fig. 16, 17, and 18. Similar relationships exist for other anion concentrations, such as halides and sulfur species, but the data base is not as well developed as the cases shown. It is important to note that these are not independent effects, because the effects of oxygen concentration and temperature are probably related to the potential. The details of this relationship are complex and are beyond the scope of this section.

Fig. 16 Effect of potential on the maximum crack growth rate in sensitized type 304 stainless steel in 0.01 m Na2SO4 at 250 °C (480 °F). Numbers denote KI values.

Fig. 17 Variation in the average crack propagation rate in sensitized type 304 stainless steel in water at 288 °C (550 °F) with oxygen content. Data are from both constant extension rate (CERT) tests, constant load, and field observations on boiling water reactor piping. IGSCC, intergranular SCC; TGSCC, transgranular SCC.

Fig. 18 Schematic of crack growth rate versus temperature for intergranular stress-corrosion cracking of type 304 stainless steel A mechanical fracture process in SCC can produce crack velocities exceeding those given by Eq 3 by some magnification factor, which can be as large as 100. The mechanism of brittle fracture induced by SCC, which is covered in the section “Crack Propagation Mechanisms” in this article, is thought to involve the formation of a corrosion product at the crack tip in which a cleavage crack can initiate and propagate some depth into the ductile substrate. As mentioned earlier, this process would produce crack lengths that exceed those accounted for by the total charge transfer/metal oxidation process, but it would be dependent on the crack tip corrosion rate because the formation rate of the corrosion product would depend on the corrosion rate. Early observations and proposals of this process have been made (Ref 21, 22, 23, 24, 25). A current list of materials in which SCC is thought to occur through a film-induced cleavage process as suggested in Ref 25 is given in Table 2, which shows a wide range of materials and environments. An attempt has been made to identify the corrosion product or layer that initiates the cleavage cracking process, but this identification must be considered speculative. The electrochemical conditions controlling the brittle SCC process have not been catalogued carefully; however, for many systems, the main issue is whether cracking can be described completely by anodic dissolution or whether a mechanical fracture process is involved. The greatest uncertainty involves the possibility of a transition between anodic and mechanical SCC processes that is affected by electrochemistry, material chemistry, or mechanics. The electrochemical factors described in the previous paragraphs generally were drawn from examples of intergranular SCC, but many of the effects would be similar for brittle transgranular SCC. For example, the brittle transgranular crack growth rate follows closely the dependence of the anodic current density on potential. Whether there are critical potentials at which transgranular SCC may switch from a brittle mechanical fracture mode to a purely anodic dissolution mode is unknown. However, for many systems, the critical potentials describing SCC are valid control parameters even though the process associated with these critical potentials may be uncertain. For an anodic dissolution-controlled process, the critical potentials were associated with the need for an active tip and passive crack walls. These conditions are still necessary for a brittle SCC process because it is dependent on the anodic reaction to produce the corrosion reaction product. This product may be a film or a dealloyed layer, but the rate of formation and, therefore, the brittle crack growth rate are dependent on the crack tip current density.

Material Chemistry and Microstructure

The relationship between material chemistry and microstructure and SCC is equally as complex as the relationship between the environment and SCC. Bulk alloy composition can affect passive film stability and phase distribution (for example, chromium in stainless steel), minor alloying elements can cause local changes in passive film-forming elements (for example, carbon in stainless steel causing sensitization), impurity elements can segregate to grain boundaries and cause local differences in the corrosion rate (for example, phosphorus in nickel or nickel-base alloys), and inclusions can cause local crack tip chemistry changes as the crack intersects them (for example, manganese sulfide in steel). In addition, alloys can undergo dealloying, which is thought to be a primary method by which brittle SCC initiates. The following summary is divided into intergranular and transgranular effects because some of the primary factors can be best described in this way; however, it is important to recognize that many of the material chemistry factors can affect both intergranular and transgranular cracking. Intergranular Stress-Corrosion Cracking. Material chemistry and microstructure effects on intergranular SCC can generally be divided into two categories: grain-boundary precipitation and grain-boundary segregation. Grain-boundary precipitation effects include carbide precipitation in austenitic stainless steels and nickel-base alloys, which causes a depletion of chromium adjacent to the grain boundary and intermetallic precipitation in aluminum alloys, which are anodically active. Grain-boundary segregation of impurities such as phosphorus, sulfur, carbon, and silicon can produce a grain boundary that is up to 50% impurity within a 1 to 2 nm thick region. These impurities can alter the corrosion and mechanical properties of the grain boundary and can therefore cause cracking by anodic dissolution and perhaps mechanical fracture. Grain-Boundary Precipitation in Stainless Steels. Chromium carbide precipitation in stainless steels occurs in the temperature range 500 to 850 °C (930 to 1560 °F), with the rate of precipitation controlled by chromium diffusion. For intermediate times, such as occurs with heat treating and welding, chromium depletion occurs adjacent to the grain boundary during chromium carbide growth. This depletion can be described by the minimum chromium concentration adjacent to the carbide and the width of the depleted zone. Minimum chromium concentrations of 8 to 10 at.% have been measured by analytical electron microscopy, while the width of the depleted zone has been measured to range from 10 nm to hundreds of nanometers. After times long enough for carbide growth to reach completion, the chromium profile is eliminated, and the chromium concentration returns to the bulk value. The intergranular SCC of austenitic stainless steel depends primarily on the nature of the chromium-depleted zone, which is explained generally by the depletion of a passive film-forming element along a continuous path through the material. The stress-corrosion susceptibility and crack growth rate of austenitic stainless steel can be described by the degree of sensitization (DOS) as measured by corrosion tests such as the Strauss or electrochemical potentiokinetic reactivation (EPR) tests. Quantitative comparisons between susceptibility as measured by the presence or absence of intergranular SCC in an SCC test or the crack growth rate and the DOS or chromium depletion parameters have been successful in cases in which sufficient data have been available, but these correlations are limited to specific alloys, environments, and stress conditions. Semiquantitative comparisons in which intergranular SCC is predicted for DOS values greater than some critical value are less costly to obtain and have been used more extensively. An example of the time/temperature/DOS curves at several carbon concentrations is shown in Fig. 19. The curves represent the conditions necessary for the development of a constant degree of sensitization, with the strong effect of bulk carbon concentration on the time needed to develop a sensitized microstructure. Therefore, the most common method for reducing the possibility of developing a sensitized microstructure is to reduce the carbon concentration or to control the thermal history of the material. Given that the material is sensitized, control of the environment and stress conditions can be used to reduce the crack propagation rate.

Fig. 19 Time/temperature/sensitization curves determined by EPR tests on type 304 stainless steel alloys of variable carbon contents Clearly, chromium and carbon are the two most important elements involved in the development of sensitization; however, other elements can play a role. Molybdenum has an effect similar to that of chromium— incorporation into the carbides and depletion adjacent to the grain boundaries—although it is generally present in smaller concentrations (for example, in type 316 stainless steel) than chromium is. Nickel increases the activity of carbon in austenitic stainless steel; therefore, increases in nickel concentration can enhance sensitization for a given carbon concentration and thermal history. Manganese, silicon, and nitrogen have also been shown to affect sensitization, but the available data on these elements are comparatively sparse. Silicon has been shown to segregate to grain boundaries when present in sufficient quantities or by the nonequilibrium processes that occur when the material is irradiated with neutrons. The presence of silicon in the grain boundaries has been found to be deleterious in oxidizing environments. Grain-Boundary Precipitation in Nickel Alloys. The cause of intergranular SCC in nickel- base alloys is more complex than it is in austenitic stainless steel. Chromium-carbide precipitation and chromium depletion occur in nickel-base alloys, such as alloy 600, as they do in austenitic stainless steels, but a clear connection between the presence of this microstructure and intergranular SCC has not been made. Carbon solubility is considerably lower and chromium diffusion is faster in nickel-base alloys than in austenitic stainless steels; therefore, the nose in the carbide precipitation or sensitization curves shown in Fig. 19 is shifted to shorter times and higher temperatures. Significant heat- to-heat variations in the chromium depletion of alloy 600 have been observed for a given heat treatment. This variation is attributed to the sensitivity of alloy 600 to carbon concentration and to the variability in mill-anneal heat treatment conditions. Chromium depletion is thought to accelerate cracking in oxygenated water but has not been identified as a controlling factor in deaerated water or caustic environments at 300 °C (570 °F). The rapid carbide precipitation kinetics in nickel-base alloys allows the development of a healed microstructure in which carbide growth is complete and the chromium profile is eliminated with relatively short heat treatment times. Some significant differences between the intergranular SCC of nickel-base alloys and that of austenitic stainless steels include the positive effect of a semicontinuous distribution of carbides at the grain boundary in singlephase material and the galvanic couple between the γ′ and γ phases in precipitation-hardened alloys. The beneficial effect of a semicontinuous distribution of carbides is thought to be related to dislocation generation from the carbides that results in a crack blunting or stress relaxation at the tip of a crack. The detrimental effect of the γ′ phase is thought to result from a galvanic couple that polarizes the γ phase anodically relative to the γ′ phase. Grain-Boundary Precipitation in Aluminum Alloys. Aluminum alloys are also noted for the occurrence of intergranular SCC in aqueous environments, including sodium chloride (NaCl) solutions. The details of the intergranular SCC process in aluminum alloys are very complex and vary with alloy composition, but some features can be summarized. Grain-boundary precipitation has been identified as a contributing factor in the intergranular SCC in aluminum alloys; galvanic effects between the precipitates and matrix are considered to be important. In some cases, the precipitates are anodic, and in others, they are cathodic to the matrix. Peak- aged materials that give maximum hardnesses and microstructures in which slip bands are produced upon straining are considered to be vulnerable to SCC. Overaged structures are considered to be less susceptible. The mechanism of crack growth is thought to be a combination of local anodic dissolution and hydrogen

embrittlement. Whether crack extension is by anodic dissolution or by hydrogen embrittlement remains an open question; however, in aqueous solutions, the primary source of hydrogen is anodic dissolution. Grain-Boundary Segregation in Aluminum Alloys. Grain-boundary enrichment of magnesium without precipitation of a magnesium-rich phase has been identified as a factor in the intergranular SCC of aluminum alloys (Ref 26, 27). This enrichment is associated with increased corrosion activity at the grain boundary and the possible formation of magnesium hydride (MgH) or enhanced hydrogen entry along the grain boundaries. Crack advance is thought to occur by discontinuous crack jumps (possibly nucleated by hydrides) and crack arrest after about 150 to 350 nm of advance, followed by repetition of this process. Grain-Boundary Segregation in Ferrous and Nickel Alloys. Grain-boundary enrichment of impurities can contribute to the intergranular SCC of iron-base alloys, austenitic stainless steels, and nickel-base alloys. The extent of their effect depends on the electrochemical potential, the presence of other stress-corrosion processes such as chromium depletion, and their concentration in the grain boundary. The enrichment of impurities to grain boundaries by equilibrium segregation can be described by the enrichment ratio, as shown in Fig. 20 for a variety of impurity- alloy combinations (Ref 28). The enrichment ratio defines the upper bound to the grainboundary concentration that will be achieved under equilibrium conditions, which can be very long times at low temperatures. It can be seen from Fig. 20 that enrichment ratios as high as 105 are possible for some impurities and that ratios greater than 102 are not uncommon. For the bulk impurity concentrations present in most engineering alloys, grain-boundary impurity concentrations of 10 to 20 at.% are possible. Therefore, an intergranular stress-corrosion crack can propagate along a grain boundary that has a composition vastly different from that of the bulk alloy.

Fig. 20 Correlation between measured grain-boundary enrichment ratios, β, and the inverse of solid solubility, . The symbols Xb and Xc represent the actual concentration of the species b and c in the bulk. Ferritic steels exhibit intergranular SCC in hot nitrate, caustic, carbonate, and a variety of other environments. The presence of intergranular SCC is dependent on the electrochemical potential because intergranular SCC is predominant at potentials in the active-passive transition. Early studies of this effect identified carbon segregation as the primary element of concern where the carbon atoms were said to provide suitable imperfection sites for adsorption of nitrate in an adsorption-induced crack growth mechanism (Ref 29). More recently, phosphorus segregation has been associated with the intergranular SCC of iron alloys in nitrate and caustic solutions (Ref 30, 31); grain-boundary concentrations as low as 2 to 3 at.% have been sufficient to alter the passivity of iron in hot nitrate solutions. A complexity of SCC in ferritic steels is the susceptibility of such steels to intergranular SCC and to hydrogen-induced subcritical crack growth. The temperatures and electrochemical potentials at which intergranular SCC and cathodic hydrogen-induced subcritical crack growth

occur generally are not the same. Stress corrosion tends to dominate at temperatures above about 50 °C (120 °F) and at potentials in the active-passive transition; hydrogen effects are predominant at temperatures below 50 °C (120 °F), at more cathodic potentials, and at lower pH values. Examples of the effect of sulfur and phosphorus on the tendency toward intergranular fracture are given in Fig. 21 and 22, from which it can be seen that sulfur was more effective than phosphorus in causing intergranular fracture of iron at cathodic potentials (Ref 32, 33). For the conditions of this test, a sulfur concentration of about 13 at.% was sufficient to change the fracture mode from transgranular to intergranular.

Fig. 21 Percent intergranular fracture, reduction of area, and strain to failure of iron, iron + phosphorus, and iron + phosphorus + manganese alloys tested at various cathodic potentials

Fig. 22 Percent intergranular fracture and the normalized strain to failure plotted as a function of sulfur content at the grain boundary for straining electrode tests at a cathodic potential of -600 mV (SCE) Grain-boundary impurity segregation of various impurities, including phosphorus, silicon, sulfur, and nitrogen, has been reported in austenitic stainless steels (Ref 34); however, no direct effect on intergranular SCC has been identified in high-temperature water. Phosphorus segregation has been shown to cause intergranular corrosion in highly oxidizing solutions, and impurity segregation of phosphorus and perhaps silicon has been suggested as a primary factor in irradiation-assisted SCC, which occurs in the oxidizing in-core environment of light water reactors (Ref 35). Phosphorus segregation apparently can contribute to the intergranular SCC of austenitic stainless steels in high-temperature water if the carbon concentration of the alloy is lower than 0.002%. At this low concentration, there is virtually no sensitization; therefore, the phosphorus segregation effect is observed. Phosphorus is also the primary grain-boundary segregant in alloy 600 and has been shown to segregate by an equilibrium process (Ref 36). The grain-boundary composition versus temperature curves for alloy 600 and type 304 stainless steel are shown in Fig. 23, along with a calculated curve for equilibrium segregation in nickel (Ref 37). It can be seen that grain-boundary phosphorus concentrations to 15 at.% are possible. Grain-boundary segregation of sulfur, boron, nitrogen, and titanium also has been observed in alloy 600; however, a clear connection between the impurity segregation and intergranular SCC in nickel-base alloys has not been observed to date.

Fig. 23 Grain-boundary segregation measurements in Alloy 600 and type 304 stainless steel. Shown are auger electron spectroscopy measurements of phosphorus segregation in the two alloys as compared with model prediction for phosphorus segregation in nickel. In contrast to the high-temperature water results for alloy 600, a clear effect of phosphorus segregation on the intergranular SCC of nickel was observed at oxidizing potentials in acidic solutions at 25 °C (77 °F) (Ref 18), as shown in Fig. 24. Intergranular SCC was observed at anodic potentials ranging from the active-passive transition to transpassive potentials in nickel with phosphorus-enriched grain boundaries but not sulfur-enriched grain boundaries. The role of phosphorus was identified with degradation of the passive film formed on nickel in acidic solutions. This example is a clear case of activepath corrosion; however, the cracking was clearly

stress dependent, as evidenced by a threshold stress intensity below which intergranular SCC did not occur and an alignment of cracks with the applied tensile stress. Because there were no grain-boundary carbides or chromium depletion in the nickel, these results illustrate the potential effect of impurities when other grainboundary processes are absent. Therefore, it appears, as in the case of type 304 stainless steel cited earlier, that impurity segregation may induce intergranular SCC in the absence of carbides and chromium depletion.

Fig. 24 Stress-corrosion cracking behavior of nickel with phosphorus and sulfur segregation. (a) Polarization curve for nickel in 1 N H2SO4 at 25 °C (77 °F). (b) Strain to failure and percent intergranular fracture for 26% phosphorus segregation at grain boundaries. (c) Strain to failure and percent intergranular fracture for 40% sulfur segregation at grain boundaries Like ferritic materials, nickel-base alloys and austenitic stainless steels are susceptible to hydrogen-induced intergranular subcritical crack growth. Impurity segregation is thought to play a key role in this hydrogeninduced fracture, as the results for nickel shown in Fig. 25 illustrate. These results show the combined effect of sulfur segregation and cathodic hydrogen on the amount of intergranular fracture in nickel. For the conditions of these tests, a given amount of intergranular fracture resulted for different combinations of grain-boundary sulfur and hydrogen reduction rate. A decrease in the amount of sulfur could be compensated for by an increased hydrogen activity to give the same amount of intergranular fracture.

Fig. 25 Percent intergranular fracture and reduction of area versus grain-boundary composition of nickel for several cathodic test potentials. CS is the critical sulfur concentration corresponding to 50% intergranular fracture. Points labeled P are equivalent sulfur concentrations for alloys with sulfur + phosphorus at the grain boundaries. Source: Ref 38 Transgranular Stress-Corrosion Cracking. Numerous metallurgical factors affect transgranular SCC, for example, crystal structure, anisotropy, grain size and shape, dislocation density and geometry, yield strength, composition, stacking fault energy, ordering, and phase composition. Some of these factors also affect intergranular SCC, as covered in the section “Intergranular Stress-Corrosion Cracking” in this article. Also, some of their effects on transgranular SCC are related to the corrosion behavior of the alloy, which can be understood from potential-pH diagrams or polarization curves. Because these effects are covered elsewhere, they are not discussed here. Similarly, yield strength effects are considered in the section “Mechanical Factors” in this article. Metallurgical factors in hydrogen-induced subcritical crack growth, such as crystal structure, hydride stability, and yield strength, are considered in the article “Hydrogen Damage” in this Volume. The effects of alloy composition on corrosion rate, hydrogen exchange current densities, and hydrogen absorption rates also influence the hydrogen-induced subcritical crack growth of materials in corrosive environments, but these are covered in the articles dealing with specific metals and alloys. Alloying effects on slip planarity are a key metallurgical factor in transgranular SCC. Planar slip occurs in alloys with low stacking fault energy, alloys containing ordered phases, or alloys exhibiting short- or longrange ordering. The consequences of planar slip on transgranular SCC have been explained by the slipdissolution model discussed in Ref 39. In this model, the passive film is ruptured by the emergence of a slip step. In high-chloride environments, evidence was presented that preferential corrosion occurred along the high dislocation density plane created by planar slip (Ref 40). A number of crack growth processes were suggested based on the planar slip localized corrosion process—for example, control of crack advance solely by anodic dissolution of the slip plane, brittle fracture of the corrosion product or tarnish film formed along the localized corrosion path, and the tunneling process by which corrosion along the slip plane branches out into tunnels accompanied by mechanical fracture of the remaining ligaments. The planar slip concept of transgranular SCC appears consistent with a number of observations, such as alloying and strain rate effects. Details of these mechanisms are presented in the section “Crack Propagation Mechanisms” in this article. A significant difficulty associated with the slip-dissolution mechanism is the nature of the transgranular SCC fracture surfaces, which are generally not on the slip planes and have cleavage-like features. For example, for

admiralty brass tested in aqueous ammonia (NH3) and Al- 5.5Zn-2.5Mg tested in aqueous NaCl, the primary fracture facets were (110) planes, while for stable austenitic stainless steel tested in aqueous magnesium chloride (MgCl2) at 155 °C (310 °F), the primary facets were on (100) planes. There is also evidence that crack advance in brass occurs in a discontinuous manner (Ref 24). Several mechanisms for the development of a cleavage crack in ductile face-centered cubic (fcc) alloys have been presented, but a definite correlation has not emerged. One concept that has received considerable attention is that proposed in Ref 25, in which a rapid crack advance that begins in a brittle film at the crack tip induces cleavage fracture of the ductile material ahead of the crack tip. The brittle films that are regarded as the initiating layers are given in Table 2. These include a dealloyed layer, oxide, nitride, carbide, or hydride. A dealloyed layer acting as a cleavage crack initiator is suggested for brass, copper-gold, and iron- chromium-nickel alloys. If this concept is supported by future research, the relative reactivity of the alloying elements will prove to be a key metallurgical factor in the transgranular SCC of materials. It has also been demonstrated that the dealloying rate of copper-aluminum and copper- zinc as a function of aluminum and zinc, respectively, corresponds to the crack growth rate in ammoniacal solution (Ref 25).

Mechanical Factors There is more commonality in the mechanical aspects of stress corrosion between intergranular SCC and transgranular SCC and between various materials than there is in environmental and metallurgical aspects. As indicated earlier, many of the environmental or metallurgical factors are specific to a given material/environment combination such that a metallurgical factor may be significant in one environment, but not in another. Threshold stress intensities and stresses, the presence of a stress-independent crack growth regime, and a dependence of crack growth rate on strain rate are features common to a variety of environmentally induced crack growth processes and a variety of materials. Anodic and cathodic controlled processes show many of these common features. The stress-intensity dependence of environmentally induced subcritical crack growth is shown schematically in Fig. 26, along with a relationship for the linear-elastic mode I stress intensity. The applied stress intensity is a function of the uniform stress and the crack length; therefore, for a constant-stress test, the stress intensity increases with increasing crack extension. For low-strength materials that form a significant plastic zone ahead of the crack, nonlinear fracture mechanics relationships must be used to calculate K. The subcritical crack growth behavior of many material/environment combinations is characterized by a threshold and by stages 1, 2, and 3 (see the article “Fracture Toughness and Fracture Mechanics” in Mechanical Testing and Evaluation, Volume 8 of ASM Handbook, for a detailed discussion of modes of crack formation).

Fig. 26 Typical, subcritical crack propagation rate versus stress intensity relationship. Stress intensity, K, is defined as are geometrical constants.

, where σ is the total tensile stress, C is the crack length, and A and B

In Fig. 26, the threshold is defined by the minimum detectable or reliably measured crack growth rate, while in some materials, stage 1 has a large K dependence such that a lower threshold does not result at lower crack velocities. There are a number of physical processes that can be associated with a threshold stress intensity, including a fracture strain for a passive film rupture mechanism, the critical resolved shear stress for the slipdissolution mechanism, a fracture stress for a brittle film-induced cleavage mechanism, or a critical crack tip opening for transport of species in the crack. The threshold stress intensity generally is associated with the development of a plastic zone at the crack tip; however, calculations of crack tip strains are not very accurate, and experimental correlations between plastic zone development of KISCC have not been made. Stage 1 subcritical crack growth is marked by a rapid increase in crack velocity with a small increase in K. Dependencies of crack velocity on stress intensity to the fourth power have been reported for a number of material/environment combinations. Changes in crack velocity of two orders of magnitude with small changes in K are not uncommon for the stage 1 regime. Few explanations for the K dependence of stage 1 have been put forth. One explanation is that in stage 1, the crack tip plastic strain rate is increasing rapidly with K, and another explanation is that the transport of species into and out of the crack increases rapidly with increasing crack volume in stage 1. In metals tested in aqueous solutions, in transport of cations or anions in the crack electrolyte, or in passive film forming materials, the passive film rupture and repassivation rates are thought to be rate-limiting processes. Environmental parameters such as crack tip potential, corrosion rates, and pH also affect the stage 2 crack velocity so that in many systems the limiting velocity is a function of crack tip corrosion rate, strain rate, and transport rate. A steady-state crack transport calculation for intergranular SCC in nickel + phosphorus gave crack growth values within a factor of five of the measured stage 2 values (Ref 16). For transgranular SCC, there is increasing evidence that the crack velocity is a function of the corrosion rate, but that crack advance results from brittle crack jumps that exceed the depth of the corrosion reaction. In this case, the crack velocity may scale with the corrosion rate, but the crack length exceeds that determined from charge transfer. Most stress-corrosion data have been obtained by using specimens loaded in a tensile mode in which the stress is perpendicular to the plane of the advancing crack; this condition is defined as mode 1 loading. In practice, a

component may have a complex loading mode such that torsion and shear components are present. The effect of mixed loading modes on SCC is therefore an important consideration in component design; however, very few SCC data exist for other than mode 1 loading. Some examples of SCC in modes 1 and 3 for aluminum in 3.5% NaCl and type 304 stainless steel in boiling MgCl2 are shown in Fig. 27, in which it can be seen that mode 1 loading gives the lowest threshold stress or stress intensity. Also shown in Fig. 27(a) is a schematic of the test specimen used to perform mixed-mode SCC tests. In both cases, the researchers concluded that the lower thresholds and higher crack velocities in mode 1 tests of these materials indicate a hydrogen embrittlement component to SCC because hydrogen effects are more prominent under tensile loading conditions (Ref 41, 42). However, anodic stress- corrosion processes are also aided by the crack opening that results from mode 1 loading.

Fig. 27 Examples of SCC in mode 1 and mode 3. (a) Susceptibility of 7075 aluminum alloy to corrosion cracking in 3.5% NaCl solution tension. Mode 1 and torsion (solid lines). Mode 3 tests (dashed lines). P,

peak and aging; OA, overaged; UA, underaged. (b) Crack velocity versus stress intensity curves for stress-corrosion cracking of type 304 stainless steel in modes 1 and 3 in boiling MgCl2 Stress-corrosion data are frequently given as applied stress versus time to failure, as shown in Fig. 28. These data are obtained with notched or unnotched specimens statically loaded in tension or bending. The threshold stress is sensitive to the environment, alloy composition, and structure and is not strictly a materials property. The threshold stress is related to the yield strength of the material, with thresholds generally being greater than one-half the yield strength, and in many cases, the threshold is a significant fraction (about 0.8) of the yield strength of the material. Therefore, most SCC tests performed with this technique are performed at values of about 80% or greater of the yield strength.

Fig. 28 Schematic diagram of time to failure versus applied stress, σ, normalized to the yield strength σy for stress corrosion The data obtained with this method incorporate the stage 1, 2, and 3 regimes shown in Fig. 26 in addition to a crack initiation stage. Crack growth begins with the development of a pit or intergranular corrosion groove that raises the stress intensity to the KISCC. Because the stress is constant, the stress intensity increases with increasing crack length as given by the expression in Fig. 26 so that the crack progresses through stage 1 to stage 2 and ultimately to fracture in stage 3. Therefore, the failure time in a constant- stress test is dependent on the initiation rate, the value of the KISCC, the crack growth rate in stage 2, and the fracture toughness KIc. These test data are useful because they relate to the sequence that may occur in practice, but it is difficult to relate them to physical processes and rate-controlling steps. Crack growth rates are frequently estimated from constant-load tests based on the crack depth divided by the total test time. Crack velocities obtained in this manner are minimum values because the time for initiation is included so that instantaneous velocities will always be greater. The threshold stress is perhaps the least complicated value obtained from constant-load tests. For film-forming materials, the threshold can be related to the film formation and rupture rate. Below the threshold stress, the film formation rate is sufficiently high to maintain the surface in a filmed condition; at higher stresses, the film rupture rate exceeds the reformation rate such that a pit or an intergranular corrosion groove can grow. For transgranular SCC initiated by a brittle corrosion film, the threshold stress could be related to the fracture stress of the film; however, this type of relationship has not been established for the brittle film-induced cleavage process. Constant extension rate or slow strain rate tests have essentially replaced the use of the constant-stress test for stress-corrosion testing. This change has occurred because of the variability in times of failure with a constantstress test and the long test times needed to define the stress threshold. Also, the concept that the crack growth rate is a function of the strain rate, as shown in Fig. 29 for intergranular type 304 stainless steel, has led to the use of a constant extension rate test. The CERT specimens will fail within a given period of time, thus eliminating specimens that do not fail because they are loaded to a stress below the threshold. The crack growth rates given in Fig. 29 are minimum values obtained from unnotched specimens, and the strain rate is the average applied rate. Clearly, once a crack develops, the crack tip strain rate will differ from the average

applied value. An example of the crack velocity as a function of the crack tip strain rate is given in Fig. 30 for intergranular SCC of a carbon-manganese steel tested in a carbonate solution. A threshold strain rate is evident in these results that is the same as a threshold stress intensity. A second threshold strain rate exists at high strain rates because mechanical fracture occurs before a crack can initiate and propagate to a measurable length. The two strain rate thresholds are illustrated in Fig. 31.

Fig. 29 Comparison between observed and theoretical crack propagation rate/strain rate ( / ) relationships for furnace-sensitized type 304 stainless steel in water/0.2 ppm oxygen at 288 °C (550 °F)

Fig. 30 Intergranular crack velocities for various applied crack tip strain rates in a carbon-manganese steel immersed in 1 N Na2CO3 + 1 N NaHCO3 at -650 mV (SCE) and 75 °C (165 °F)

Fig. 31 Stress-corrosion crack growth as a function of the two strain rate thresholds,

1

and

2

It is important to distinguish between static and dynamic effects. For example, a sample could be plastically strained 10% and the load dropped to below the threshold stress before being immersed in a corrosive environment and SCC will not occur. However, a sample can be slowly strained 10% in a corrosive environment with the development of numerous cracks and perhaps even complete failure. Therefore, it is the dynamic strain and creation of fresh surface or dynamic fracture of a brittle film that is crucial to SCC. Corrosion current measurements corroborate these observations; the current density of a prestrained sample is usually very similar to an unstrained sample, while the current density measured during dynamic strain is considerably greater than that in the unstrained sample.

References cited in this section 16. M.J. Danielson, C. Oster, and R.H. Jones, Corros. Sci., Vol 32 (No 1), 1991, p 1–21 18. A. Turnbull, Progress in the Understanding of the Electrochemistry in Cracks, Embrittlement by the Local Crack Environment, R.P. Gangloff, Ed., The Metallurgical Society, 1984, p 3 19. R.N. Parkins, Prevention of Environment Sensitive Fracture by Inhibition, Embrittlement by the Local Crack Environment, R.P. Gangloff, Ed., The Metallurgical Society, 1984, p 385 20. R.N. Parkins, Br. Corros. J., Vol 14, 1979, p 5 21. C. Edeleneau and A.J. Forty, Philos. Mag., Vol 46, 1960, p 521 22. J.A. Beavers and E.N. Pugh, Metall. Trans. A, Vol 11A, 1980, p 809 23. M.T. Hahn and E.N. Pugh, Corrosion, Vol 36, 1980, p 380 24. E.N. Pugh, On the Propagation of Transgranular Stress Corrosion Cracks, Atomistics of Fracture, R.M. Latanision and J.R. Pickens, Ed., Plenum Press, 1983, p 997 25. R.C. Newman and K. Sieradzki, “Film-Induced Cleavage During Stress-Corrosion Cracking of Ductile Metals and Alloys,” NATO Advanced Research Workshop on Chemistry and Physics of Fracture, June 1986 26. N.J.H. Holroyd and G.M. Scamans, Scr. Metall., Vol 19, 1985, p 915 27. E.C. Pow, W.W. Gerberich, and L.E. Toth, Scr. Metall., Vol 15, 1981, p 55 28. E.D. Hondros and M.P. Seah, Int. Met. Rev., Dec 1977, p 262

29. L. Long and H. Uhlig, J. Electrochem. Soc., Vol 112, 1965 30. J. Kuppa, H. Erhart, and H. Grabke, Corros. Sci., Vol 21, 1981, p 227 31. N. Bandyopadhyay, R.C. Newman, and K. Sieradzdki, in Proceedings of the Ninth International Congress on Metallic Corrosion (Toronto, Canada), NACE, 1984 32. R.H. Jones, S.M. Bruemmer, M.T. Thomas, and D.R. Baer, Comparison of Segregated Phosphorus and Sulfur Effects on the Fracture Mode and Ductility of Iron Tested at Cathodic Potentials, Scr. Metall., Vol 16, 1982, p 615 33. S.M. Bruemmer, R.H. Jones, M.T. Thomas, and D.R. Baer, Fracture Mode Transition of Iron in Hydrogen as a Function of Grain Boundary Sulfur, Scr. Metall., Vol 14, 1980, p 137 34. A. Joshi and D.J. Stein, Corrosion, Vol 28 (No. 9), 1972, p 321 35. R.H. Jones, Some Radiation Damage-Stress Corrosion Synergisms in Austenitic Stainless Steels, Proceedings of the Second International Symposium on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactors, American Nuclear Society, Sept 1985, p 173 36. M. Guttman, P. Dumoulin, Nguyen Tan Tai, and P. Fontaine, Corrosion, Vol 37, 1981, p 416 37. S.M. Bruemmer, L.A. Charlot, and C.H. Henager, Jr., “Microstructural Effects on Microdeformation and Primary-Side Stress Corrosion Cracking of Alloy 600 Tubing,” EPRI Final Report, Project 2163-4, Electric Power Research Institute, Aug 1986 38. R.H. Jones, S.M. Bruemmer, M.T. Thomas, and D.R. Baer, Influence of Sulfur, Phosphorus, and Antimony Segregation on the Intergranular Hydrogen Embrittlement of Nickel, Metall. Trans. A, Vol 14A, 1983, p 223 39. R.W. Staehle et al., Corrosion, Vol 26 (No. 11), 1970, p 451 40. H.W. Pickering and P.R. Swann, Corrosion, Vol 19 (No. 3), 1963, p 373 41. A.W. Thompson and I.M. Bernstein, Advances in Corrosion Sciences and Technology, Vol 7, M.G. Fontana and R.W. Staehle, Ed., Plenum Press, 1980, p 53 42. R.M. Riecke, A. Athens, and I.O. Smith, Mater. Sci. Technol., Vol 2, 1986, p 1066


Crack Propagation Mechanisms As indicated in the previous discussion, it is unlikely that a single mechanism of SCC exists; instead, there are probably two or three different operative mechanisms. Many models have been proposed, but only the more

prominent are discussed. The proposed mechanisms for crack propagation fall into two basic classifications: those based on dissolution as the principal cause of crack propagation and those that involve mechanical fracture in the crack propagation process.

Dissolution Models According to models of this type, the crack advances by preferential dissolution at the crack tip. A number of models have been proposed to account for this process. For example, preferential dissolution at the crack tip has been attributed to the formation of active paths in the material, stresses at the crack tip, and chemicalmechanical interactions (Ref 43, 44, 45, 46, 47, 48). However, research has essentially eliminated all but one model from serious consideration, and the debate has recently centered on the details of this model. Film Rupture. The film rupture model, also referred to as the slip-dissolution model, assumes that the stress acts to open the crack and rupture the protective surface film. Two investigators, working independently, first postulated that localized plastic deformation at the crack tip ruptures the passivating film, exposing bare metal at the crack tip (Ref 49, 50). The freshly exposed bare metal then dissolves rapidly, resulting in crack extension. Some investigators (Ref 51, 52, 53) assume that once propagation starts, the crack tip remains bare because the rate of film rupture at the crack tip is greater than the rate of repassivation (Fig. 32a). Others (Ref 54, 55, 56, 57) assume that the crack tip repassivates completely and is ruptured periodically by the emergence of slip steps (Fig. 32b).

Fig. 32 Schematic representations of crack propagation by the film rupture model. (a) Ref 50. (b) Ref 56 and 57 Considerable evidence has been found to support these mechanistic models, and intergranular corrosion may be considered the low stress limiting case of this mechanism (Ref 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57). However, the observation of discontinuous cracking and crack arrest markings (discussed previously) is an indication that crack propagation can be, and frequently is, discontinuous. Also, transgranular SCC fracture surfaces are flat, crystallographically oriented, and match precisely on opposite sides of the fracture surface (indicating very little dissolution during crack advance). As a result, film rupture and dissolution are accepted as viable mechanisms of intergranular SCC in some systems, but generally are not accepted as mechanisms of transgranular SCC (Ref 58).

Mechanical Fracture Models Mechanical fracture models originally assumed that stress concentration at the base of corrosion slots or pits increased to the point of ductile deformation and fracture (Ref 48). These early proposals assumed that the crack essentially propagated by dissolution and that the remaining ligaments then failed by mechanical fracture (ductile or brittle). A refinement of this approach has been proposed (Ref 40, 59); it is generally known as the corrosion tunnel model. The Corrosion Tunnel Model. In this model, it is assumed that a fine array of small corrosion tunnels form at emerging slip steps. These tunnels grow in diameter and length until the stress in the remaining ligaments

causes ductile deformation and fracture. The crack thus propagates by alternating tunnel growth and ductile fracture (Fig. 33). Cracks propagating by this mechanism should result in grooved fracture surfaces with evidence of microvoid coalescence on the peaks, as illustrated in Fig. 33(a). This theory is not consistent with the common fractographic features discussed previously. As a result, it was suggested (Ref 60) that the application of a tensile stress results in a change in the morphology of the corrosion damage from tunnels to thin flat slots, as shown in Fig. 33(b). The formation of slots was observed below the dealloyed sponge layer on the {110} type planes in austenitic stainless steels, and it was found that the formation of these slots required the presence of a tensile stress. The width of the corrosion slots was found to approach atomic dimensions, and as a result, close correspondence of matching fracture surfaces would be expected. It was concluded that transgranular SCC can be explained in terms of the formation and mechanical separation of corrosion slots (Ref 60).

Fig. 33 Corrosion tunnel models. (a) Schematic of tunnel model showing the initiation of a crack by the formation of corrosion tunnels at slip steps and ductile deformation and fracture of the remaining ligaments. (b) Schematic diagram of the tunnel mechanism of SCC and flat slot formation as proposed in Ref 60 Adsorption-Enhanced Plasticity. Studies of liquid-metal embrittlement (LME), hydrogen embrittlement, and SCC have led to the conclusion that similar fracture processes occur in each case (Ref 61, 62, 63). Because chemisorption is common to all three, this process was concluded to be responsible for the environmentally induced crack propagation (Ref 61). Based on fractographic studies, it was concluded that cleavage fracture is not an atomically brittle process but occurs instead by alternate slip at the crack tip in conjunction with formation of very small voids ahead of the crack. It was also proposed that chemisorption of environmental species facilitated the nucleation of dislocations at the crack tip, promoting the shear processes responsible for brittle cleavagelike fracture (Ref 61, 62, 63). The origin of crack arrest markings by this mechanism is, however, uncertain. Nevertheless, this mechanism promises to explain many similarities among SCC, LME, and hydrogen embrittlement. The Tarnish Rupture Model. This model was first proposed (Ref 64) to explain transgranular SCC but was later modified by other researchers (Ref 65, 66, 67) to explain intergranular SCC. In the original model (Ref 64), a brittle surface film forms on the metal that fractures under the applied stress. Fracture of the film exposes bare metal, which rapidly reacts with the environment to re-form the surface film. The crack propagates by alternating film growth and fracture, as shown in Fig. 34(a). This hypothesis was later modified to explain intergranular SCC based on the assumption that the oxide film penetrates along the grain boundary ahead of the crack tip, as shown in Fig. 34(b). Again, crack propagation consists of alternating periods of film growth and brittle film fracture. Film growth requires transport of species across the film and, as a result, the thickness of the film is limited in the absence of stress (Ref 67). This model predicts crack arrest markings on intergranular fracture surfaces and discontinuous acoustic emission during crack propagation that are not always observed during intergranular SCC. Also, it assumed penetration of the film into the grain boundary ahead of the crack tip, which may not be the case for all systems (Ref 68). At present, experimental results are insufficient to confirm or refute this model (Ref 66, 67).

Fig. 34 Tarnish rupture models. (a) Schematic of the tarnish rupture model for SCC as proposed in Ref 64. (b) Modified tarnish rupture model of SCC for systems with intergranular oxide film penetration (Ref 66, 67) The Film-Induced Cleavage Model. In 1959, the hypothesis was presented that dealloying and/or vacancy injection could induce brittle fracture (Ref 69). However, the exact nature of the interactions and how they could induce the observed crack propagation was not thoroughly evaluated. More recently, a model was developed based on the hypothesis that a surface film could induce cleavage fracture (Ref 70). This model assumes that: • • • • •

A thin surface film or layer forms on the surface A brittle crack initiates in this layer The brittle crack crosses the film/matrix interface with little loss in velocity Once in the ductile matrix, the brittle crack will continue to propagate This crack will eventually blunt and arrest, after which this process repeats itself

This model has the unique ability to explain the crack arrest markings, the cleavagelike facets on the fracture surface, and the discontinuous nature of crack propagation. The hypothesis that a brittle crack will continue to propagate after it has entered the normally ductile matrix is a critical point. This allows a thin surface layer to induce brittle crack propagation over distances much greater than the film thickness. A critical examination of this hypothesis concluded that a brittle crack can propagate in a ductile matrix if the crack is sharp and is propagating at high velocities before entering the ductile matrix (Ref 71). A computer model was developed for this process, and it was concluded that a surface layer can initiate brittle fracture even if the layer is ductile (depending on lattice mismatch, etc.) (Ref 72). More research into surface films and brittle fracture is required before this model can be thoroughly evaluated. Adsorption-Induced Brittle Fracture. This model, which was initially presented in Ref 46, is based on the hypothesis that adsorption of environmental species lowers the interatomic bond strength and the stress required for cleavage fracture. This model is frequently referred to as the stress-sorption model, and similar mechanisms have been proposed for hydrogen embrittlement and LME (Ref 73). This model predicts that cracks should propagate in a continuous manner at a rate determined by the arrival of the embrittling species at the crack tip. This model does not explain how the crack maintains an atomically sharp tip in a normally ductile material, because it does not include a provision for limiting deformation in the plastic zone. Also, the discontinuous nature of crack propagation is not explained by this model. Hydrogen Embrittlement. Stress-corrosion cracking in some material/environment combinations can be a form of hydrogen-induced subcritical crack growth. Because the anodic reaction must have a corresponding cathodic reaction and because the reduction of hydrogen is frequently the cathodic reaction, hydrogen-induced subcritical crack growth can be the dominant stress-corrosion crack growth process in some materials. Many features of hydrogen-induced subcritical crack growth from cathodic hydrogen are very similar to those produced by gaseous or internal hydrogen. The mechanisms of hydrogen damage are discussed in the following section. However, a few features of hydrogen embrittlement from cathodic hydrogen that are different from other forms of hydrogen embrittlement are summarized. Once hydrogen has been absorbed by a material, its effect, whether from a gaseous or cathodic source, is the same. This has been shown for a number of materials and a variety of properties. There are three primary differences between gaseous and cathodic hydrogen absorption processes. First, cathodic hydrogen adsorbs on the surface as atomic hydrogen (as reduced), while gaseous hydrogen adsorbs in the molecular form and must dissociate to form atomic hydrogen. Desorption of loosely bound molecular hydrogen is relatively easy, while the dissociation step can, in some cases, be the rate- determining step. Therefore, the desorption and absorption rates of gaseous and cathodic hydrogen may be substantially different for equal hydrogen activities. Second, the hydrogen activities produced by cathodic hydrogen can be quite large (thousands of psi) and are dependent on the anodic reaction rate, while gaseous hydrogen pressures are generally much lower. Lastly, the surface of the material at a crack tip may be substantially different under electrochemical corrosion conditions from that in the presence of gaseous hydrogen containing substantial quantities of other gases, such a O2 and CO2. Hydrogen-induced crack growth as the dominant stress-corrosion mechanism has been suggested for ferritic steels, nickel-base alloys, titanium alloys, and aluminum alloys, although ferritic steels show the most evidence

of this mechanism. The effects of factors such as yield strength, impurity segregation, and temperature on the crack growth behavior of ferritic materials in aqueous environments all follow the trends of gaseous hydrogen embrittlement. For example, the crack growth rate of a 3% Ni steel as a function of temperature in water is shown in Fig. 35(a), while the crack growth rate of 4340 steel in gaseous hydrogen is shown in Fig. 35(b). The maximum crack growth rate occurs at about 20 °C (70 °F), with similar decreases at both higher and lower temperatures. However, anodic stress- corrosion processes become active at temperatures above 100 °C (212 °F) for the steel tested in water. Similar trends have been shown for impurity segregation effects on hydrogeninduced crack growth of materials where cathodic and gaseous hydrogen produce essentially similar results. Some of these results have been presented in the section “Material Chemistry and Microstructure” in this article.

Fig. 35 Schematic of crack growth rate versus temperature for (a) 3% Ni steel in water (Ref 74) and (b) 4340 steel in gaseous hydrogen (Ref 75) Specific mechanisms of cathodic hydrogen induced subcritical crack growth have not been developed because it has generally been sufficient merely to identify hydrogen as the cause for cracking. However, a mechanism was presented in which grain-boundary impurities act as hydrogen recombinant poisons and enhance the uptake of cathodic hydrogen (Ref 76). A schematic of this process is shown in Fig. 36, in which the tin and antimony that segregated to the grain boundaries of nickel enhance the hydrogen uptake kinetics. This mechanism does not propose a new mechanism by which hydrogen can cause cracking but merely proposes a mechanism by which impurity segregation can enhance hydrogen uptake. There may be circumstances in which such a mechanism could tip the balance between an anodically driven process and a cathodically driven process or could accelerate cracking to a value that is measurable or of practical significance. However, a review of the combined effects of impurity segregation and hydrogen embrittlement concluded that grain- boundary impurities behave the same with cathodic and gaseous hydrogen in that they enhance crack growth by a combined grain-boundary embrittlement processes but not by enhanced hydrogen uptake (Ref 77). It should be

noted, however, that a material with impurities such as sulfur, phosphorus, antimony, and tin segregated to their grain boundaries are more susceptible to all forms of hydrogen—cathodic, gaseous, or internal.

Fig. 36 Schematic showing effect of some impurities on mechanism by which intergranular embrittlement of nickel is presumed to occur at cathodic potentials

References cited in this section 40. H.W. Pickering and P.R. Swann, Corrosion, Vol 19 (No. 3), 1963, p 373 43. E.H. Dix, Trans. AIME, Vol 137 (No 11), 1940 44. R.B. Mears, R.H. Brown, and E.H. Dix, Symposium on Stress-Corrosion Cracking of Metals, ASTM and the American Institute of Mining, Metallurgical, and Petroleum Engineers, 1944, p 323 45. E.H. Dix, Jr., Trans. ASM, Vol 42, 1950, p 1057 46. H.H. Uhlig, Physical Metallurgy of Stress Corrosion Fracture, T.N. Rhodin, Ed., Interscience, 1959, p 1 47. L. Yang, G.T. Horne, and G.M. Pound, Physical Metallurgy of Stress Corrosion Fracture, T.N. Rhodin, Ed., Interscience, 1950, p 29 48. J.J. Harwood, Stress Corrosion Cracking and Embrittlement, W.D. Robertson, Ed., John Wiley & Sons, 1956, p 1 49. F.A. Champion, in Symposium on Internal Stresses in Metals and Alloys, Institute of Metals, 1948, p 468 50. H.L. Logan, J. Res. Natl. Bur. Stand., Vol 48, 1952, p 99

51. E.W. Hart, Surfaces and Interfaces II, Syracuse University Press, 1968, p 210 52. H.J. Engle, in The Theory of Stress Corrosion Cracking in Alloys, North Atlantic Treaty Organization, 1971 53. J.C. Scully, Corros. Sci., Vol 15, 1975, p 207 54. D.A. Vermilyea, J. Electrochem. Soc., Vol 119, 1972, p 405 55. D.A. Vermilyea, Stress Corrosion Cracking and Hydrogen Embrittlement of Iron Based Alloys, NACE, 1977, p 208 56. R.W. Staehle, in The Theory of Stress Corrosion Cracking in Alloys, North Atlantic Treaty Organization, 1971, p 233 57. R.W. Staehle, Stress Corrosion Cracking and Hydrogen Embrittlement of Iron Based Alloys, NACE, 1977, p 180 58. E.N. Pugh, Corrosion, Vol 41 (No. 9), 1985, p 517 59. P.R. Swann and J.D. Embury, High Strength Materials, John Wiley & Sons, 1965, p 327 60. J.M. Silcock and P.R. Swann, Environment- Sensitive Fracture of Engineering Materials, Z.A. Foroulis, Ed., The Metallurgical Society, 1979, p 133 61. S.P. Lynch, Hydrogen Effects in Metals, A.W. Thompson and I.M. Bernstein, Ed., The Metallurgical Society, 1981, p 863 62. S.P. Lynch, Met. Sci., Vol 15 (No. 10), 1981, p 463 63. S.P. Lynch, J. Mater. Sci., Vol 20, 1985, p 3329 64. A.J. Forty and P. Humble, Philos. Mag., Vol 8, 1963, p 247 65. A.J. McEvily and P.A. Bond, J. Electrochem. Soc., Vol 112, 1965, p 141 66. E.N. Pugh, in Stress Corrosion Cracking and Hydrogen Embrittlement of Iron Based Alloys, National Association of Corrosion Engineers, 1977, p 37 67. J.A. Beavers, I.C. Rosenberg, and E.N. Pugh, in Proceedings of the 1972 Tri-Service Conference on Corrosion, MCIC-73- 19, Metals and Ceramics Information Center, 1972, p 57 68. T.R. Pinchback, S.P. Clough, and L.A. Heldt, Metall, Trans. A, Vol 7A, 1976, p 1241; Metall. Trans. A, Vol 6A, 1975, p 1479 69. A.J. Forty, Physical Metallurgy of Stress Corrosion Fracture, T.N. Rhodin, Ed., Interscience, 1959, p 99 70. K. Sieradzki and R.C. Newman, Philos. Mag. A, Vol 51 (No. 1), 1985, p 95 71. I.-H. Lin and R.M. Thomson, J. Mater. Res., Vol 1 (No. 1), 1986 72. G.J. Dienes, K. Sieradzki, A. Paskin, and B. Massoumzadeh, Surf. Sci., Vol 144, 1984, p 273

73. N.S. Stolloff, Environment-Sensitive Fracture of Engineering Materials, Z.A. Foroulis, Ed., The Metallurgical Society, 1979, p 486 74. M.O. Speidel and R.M. Magdowski, Stress Corrosion Cracking of Steam Turbine Steels—An Overview, Proceedings of the Second International Symposium on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactors, American Nuclear Society, 1986, p 267 75. D.P. Williams and H.G. Nelson, Metall. Trans., Vol 1, 1970, p 63 76. R.M. Latanision and H. Opperhauser, Metall. Trans., Vol 5, Scientific and Technical Book Service, 1974, p 483 77. R.H. Jones, A Review of Combined Impurity Segregation-Hydrogen Embrittlement Processes, Advances in the Mechanics and Physics of Surfaces, R.M. Latanision and T.E. Fischer, Ed., Scientific and Technical Book Service, 1986


Summary Stress-corrosion cracking is a phenomenon in which time-dependent crack growth occurs when the necessary electrochemical, mechanical, and metallurgical conditions exist. Corrosion fatigue is a related process in which the load is cyclic in corrosion fatigue rather than static as in stress corrosion. When hydrogen is generated as a product of the corrosion reaction, crack growth can occur by a hydrogen embrittlement process in much the same way as if hydrogen were in the gaseous form. A common feature of each of these processes is the subcritical crack growth in which cracks grow from existing flaws or initiation sites and grow to a size at which catastrophic failure occurs. Catastrophic failure occurs because the combination of crack length and applied stress increases the stress intensity to the fracture toughness of the material. A second common feature of stress corrosion, corrosion fatigue, and hydrogen-induced crack growth is that these mechanisms do not require the entire component to become embrittled; the effect is localized to the crack tip region. This article is intended to familiarize the reader with the phenomenological and mechanistic aspects of stress corrosion in order to enhance the use of other discussions in this Handbook on stress-corrosion evaluation and occurrence in specific industries and environments. The phenomenological description of crack initiation and propagation describes well-established experimental evidence and observations of stress corrosion, while the discussions on mechanisms describe the physical process involved in crack initiation and propagation. The physical processes involved in crack growth have received more evaluation and are better understood than the processes responsible for crack initiation. There is some phenomenological understanding of stress-corrosion crack initiation, but the detailed mechanistic information is not well known. Stress-corrosion cracking occurs when certain critical conditions are achieved. These conditions include electrochemical, mechanical, and metallurgical factors that must exist simultaneously. A change in any one of these three factors is adequate for eliminating SCC; therefore, a clear knowledge of these critical factors is important in system design. The important electrochemical parameters include oxidizing potential, pH, impurity concentration, and temperature. The important mechanical parameters include stress, stress intensity, and strain rate. The important metallurgical factors include localized micro-chemistry (such as the depletion of passive

film forming elements or the enrichment of active corroding elements), bulk composition, deformation character, and yield strength. There are a number of plausible mechanisms or physical processes that account for SCC. No single mechanism is adequate to describe stress corrosion in the variety of materials in which it has been observed. Stresscorrosion crack propagation mechanisms can be subdivided into dissolution and mechanical fracture models, with the mechanical fracture models further divided into ductile and brittle crack extension processes. Dissolution models include film-rupture and active-path processes, while ductile mechanical models include corrosion tunneling and adsorption-enhanced plasticity models. Brittle mechanical models include the tarnish rupture and film- induced cleavage models.


Acknowledgment This article is adapted from the article “Stress- Corrosion Cracking,” by R.H. Jones and R.E. Ricker, in the previous edition, Corrosion, Volume 13 of the ASM Handbook, ASM International, 1987.


References 1. T.A. Michalske and S.W. Freiman, Nature, Vol 295, 11 Feb 1982, p 511 2. J.F. Fellers and B.F. Kee, J. Appl. Polymer Sci., Vol 18, 1974, p 2355 3. E.H. Andrews, Developments in Polymer Science, Applied Science, 1979 4. H.T. Sumison and D.P. Williams, in Fatigue of Composite Materials, STP 569, ASTM, 1975, p 226 5. C. Zeben, in Analysis of the Test Methods for High Modulus Fibers and Composites, STP 521, ASTM, 1973, p 65 6. K. Sieradski and R.C. Newman, Philos. Mag. A, Vol 5 (No. 1), 1985, p 95 7. T. Cassaigne, E.N. Pugh, and J. Kruger, National Bureau of Standards and Johns Hopkins University, unpublished research, 1987

8. C.M. Chen, M.H. Froning, and E.D. Verink, in Stress Corrosion—New Approaches, STP 610, ASTM, 1976, p 289 9. C.J. Beevers, Ed., The Measurement of Crack Length and Shape During Fracture and Fatigue, Chameleon Press, 1980 10. E.N. Pugh, U. Bertocci, and R.E. Ricker, National Bureau of Standards, unpublished research, 1987 11. D.P. Williams and H.G. Nelson, Metall. Trans., Vol 1 (No. 1), 1970, p 63 12. R.E. Ricker and D.J. Duquette, The Role of Environment on Time Dependent Crack Growth, Micro and Macro Mechanics of Crack Growth, K. Sadananda, B.B. Rath, and D.J. Michel, Ed., American Institute of Mining, Metallurgical, and Petroleum Engineers, 1982, p 29 13. R.M. Parkins, Mater. Sci. Technol., Vol 1, 1985, p 480 14. L. Hagn, Lifetime Prediction for Parts in Corrosion Environments, Corrosion in Power Generating Equipment, Plenum Press, 1983 15. O. Buck and R. Ranjan, Evaluation of a Crack-Tip-Opening Displacement Model Under StressCorrosion Conditions, Modeling Environmental Effects on Crack Growth Processes, R.H. Jones and W.W. Gerberich, Ed., The Metallurgical Society, 1986, p 209 16. M.J. Danielson, C. Oster, and R.H. Jones, Corros. Sci., Vol 32 (No 1), 1991, p 1–21 17. R.H. Jones, M.J. Danielson, and D.R. Baer, J. Mater. Energy Syst., Vol 8, 1986, p 185 18. A. Turnbull, Progress in the Understanding of the Electrochemistry in Cracks, Embrittlement by the Local Crack Environment, R.P. Gangloff, Ed., The Metallurgical Society, 1984, p 3 19. R.N. Parkins, Prevention of Environment Sensitive Fracture by Inhibition, Embrittlement by the Local Crack Environment, R.P. Gangloff, Ed., The Metallurgical Society, 1984, p 385 20. R.N. Parkins, Br. Corros. J., Vol 14, 1979, p 5 21. C. Edeleneau and A.J. Forty, Philos. Mag., Vol 46, 1960, p 521 22. J.A. Beavers and E.N. Pugh, Metall. Trans. A, Vol 11A, 1980, p 809 23. M.T. Hahn and E.N. Pugh, Corrosion, Vol 36, 1980, p 380 24. E.N. Pugh, On the Propagation of Transgranular Stress Corrosion Cracks, Atomistics of Fracture, R.M. Latanision and J.R. Pickens, Ed., Plenum Press, 1983, p 997 25. R.C. Newman and K. Sieradzki, “Film-Induced Cleavage During Stress-Corrosion Cracking of Ductile Metals and Alloys,” NATO Advanced Research Workshop on Chemistry and Physics of Fracture, June 1986 26. N.J.H. Holroyd and G.M. Scamans, Scr. Metall., Vol 19, 1985, p 915 27. E.C. Pow, W.W. Gerberich, and L.E. Toth, Scr. Metall., Vol 15, 1981, p 55 28. E.D. Hondros and M.P. Seah, Int. Met. Rev., Dec 1977, p 262

29. L. Long and H. Uhlig, J. Electrochem. Soc., Vol 112, 1965 30. J. Kuppa, H. Erhart, and H. Grabke, Corros. Sci., Vol 21, 1981, p 227 31. N. Bandyopadhyay, R.C. Newman, and K. Sieradzdki, in Proceedings of the Ninth International Congress on Metallic Corrosion (Toronto, Canada), NACE, 1984 32. R.H. Jones, S.M. Bruemmer, M.T. Thomas, and D.R. Baer, Comparison of Segregated Phosphorus and Sulfur Effects on the Fracture Mode and Ductility of Iron Tested at Cathodic Potentials, Scr. Metall., Vol 16, 1982, p 615 33. S.M. Bruemmer, R.H. Jones, M.T. Thomas, and D.R. Baer, Fracture Mode Transition of Iron in Hydrogen as a Function of Grain Boundary Sulfur, Scr. Metall., Vol 14, 1980, p 137 34. A. Joshi and D.J. Stein, Corrosion, Vol 28 (No. 9), 1972, p 321 35. R.H. Jones, Some Radiation Damage-Stress Corrosion Synergisms in Austenitic Stainless Steels, Proceedings of the Second International Symposium on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactors, American Nuclear Society, Sept 1985, p 173 36. M. Guttman, P. Dumoulin, Nguyen Tan Tai, and P. Fontaine, Corrosion, Vol 37, 1981, p 416 37. S.M. Bruemmer, L.A. Charlot, and C.H. Henager, Jr., “Microstructural Effects on Microdeformation and Primary-Side Stress Corrosion Cracking of Alloy 600 Tubing,” EPRI Final Report, Project 2163-4, Electric Power Research Institute, Aug 1986 38. R.H. Jones, S.M. Bruemmer, M.T. Thomas, and D.R. Baer, Influence of Sulfur, Phosphorus, and Antimony Segregation on the Intergranular Hydrogen Embrittlement of Nickel, Metall. Trans. A, Vol 14A, 1983, p 223 39. R.W. Staehle et al., Corrosion, Vol 26 (No. 11), 1970, p 451 40. H.W. Pickering and P.R. Swann, Corrosion, Vol 19 (No. 3), 1963, p 373 41. A.W. Thompson and I.M. Bernstein, Advances in Corrosion Sciences and Technology, Vol 7, M.G. Fontana and R.W. Staehle, Ed., Plenum Press, 1980, p 53 42. R.M. Riecke, A. Athens, and I.O. Smith, Mater. Sci. Technol., Vol 2, 1986, p 1066 43. E.H. Dix, Trans. AIME, Vol 137 (No 11), 1940 44. R.B. Mears, R.H. Brown, and E.H. Dix, Symposium on Stress-Corrosion Cracking of Metals, ASTM and the American Institute of Mining, Metallurgical, and Petroleum Engineers, 1944, p 323 45. E.H. Dix, Jr., Trans. ASM, Vol 42, 1950, p 1057 46. H.H. Uhlig, Physical Metallurgy of Stress Corrosion Fracture, T.N. Rhodin, Ed., Interscience, 1959, p 1 47. L. Yang, G.T. Horne, and G.M. Pound, Physical Metallurgy of Stress Corrosion Fracture, T.N. Rhodin, Ed., Interscience, 1950, p 29 48. J.J. Harwood, Stress Corrosion Cracking and Embrittlement, W.D. Robertson, Ed., John Wiley & Sons, 1956, p 1

49. F.A. Champion, in Symposium on Internal Stresses in Metals and Alloys, Institute of Metals, 1948, p 468 50. H.L. Logan, J. Res. Natl. Bur. Stand., Vol 48, 1952, p 99 51. E.W. Hart, Surfaces and Interfaces II, Syracuse University Press, 1968, p 210 52. H.J. Engle, in The Theory of Stress Corrosion Cracking in Alloys, North Atlantic Treaty Organization, 1971 53. J.C. Scully, Corros. Sci., Vol 15, 1975, p 207 54. D.A. Vermilyea, J. Electrochem. Soc., Vol 119, 1972, p 405 55. D.A. Vermilyea, Stress Corrosion Cracking and Hydrogen Embrittlement of Iron Based Alloys, NACE, 1977, p 208 56. R.W. Staehle, in The Theory of Stress Corrosion Cracking in Alloys, North Atlantic Treaty Organization, 1971, p 233 57. R.W. Staehle, Stress Corrosion Cracking and Hydrogen Embrittlement of Iron Based Alloys, NACE, 1977, p 180 58. E.N. Pugh, Corrosion, Vol 41 (No. 9), 1985, p 517 59. P.R. Swann and J.D. Embury, High Strength Materials, John Wiley & Sons, 1965, p 327 60. J.M. Silcock and P.R. Swann, Environment- Sensitive Fracture of Engineering Materials, Z.A. Foroulis, Ed., The Metallurgical Society, 1979, p 133 61. S.P. Lynch, Hydrogen Effects in Metals, A.W. Thompson and I.M. Bernstein, Ed., The Metallurgical Society, 1981, p 863 62. S.P. Lynch, Met. Sci., Vol 15 (No. 10), 1981, p 463 63. S.P. Lynch, J. Mater. Sci., Vol 20, 1985, p 3329 64. A.J. Forty and P. Humble, Philos. Mag., Vol 8, 1963, p 247 65. A.J. McEvily and P.A. Bond, J. Electrochem. Soc., Vol 112, 1965, p 141 66. E.N. Pugh, in Stress Corrosion Cracking and Hydrogen Embrittlement of Iron Based Alloys, National Association of Corrosion Engineers, 1977, p 37 67. J.A. Beavers, I.C. Rosenberg, and E.N. Pugh, in Proceedings of the 1972 Tri-Service Conference on Corrosion, MCIC-73- 19, Metals and Ceramics Information Center, 1972, p 57 68. T.R. Pinchback, S.P. Clough, and L.A. Heldt, Metall, Trans. A, Vol 7A, 1976, p 1241; Metall. Trans. A, Vol 6A, 1975, p 1479 69. A.J. Forty, Physical Metallurgy of Stress Corrosion Fracture, T.N. Rhodin, Ed., Interscience, 1959, p 99 70. K. Sieradzki and R.C. Newman, Philos. Mag. A, Vol 51 (No. 1), 1985, p 95

71. I.-H. Lin and R.M. Thomson, J. Mater. Res., Vol 1 (No. 1), 1986 72. G.J. Dienes, K. Sieradzki, A. Paskin, and B. Massoumzadeh, Surf. Sci., Vol 144, 1984, p 273 73. N.S. Stolloff, Environment-Sensitive Fracture of Engineering Materials, Z.A. Foroulis, Ed., The Metallurgical Society, 1979, p 486 74. M.O. Speidel and R.M. Magdowski, Stress Corrosion Cracking of Steam Turbine Steels—An Overview, Proceedings of the Second International Symposium on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactors, American Nuclear Society, 1986, p 267 75. D.P. Williams and H.G. Nelson, Metall. Trans., Vol 1, 1970, p 63 76. R.M. Latanision and H. Opperhauser, Metall. Trans., Vol 5, Scientific and Technical Book Service, 1974, p 483 77. R.H. Jones, A Review of Combined Impurity Segregation-Hydrogen Embrittlement Processes, Advances in the Mechanics and Physics of Surfaces, R.M. Latanision and T.E. Fischer, Ed., Scientific and Technical Book Service, 1986


Selected References General Reference • • • • • •

R.P. Gangloff and M.B. Ives, Ed., “Environment-Induced Cracking of Metals,” NACE- 10, NACE, 1988 R.H. Jones, Ed., Chemistry and Electrochemistry of Corrosion and Stress Corrosion Cracking: A Symposium Honoring the Contributions of R.W. Staehle, TMS, 2001 R.H. Jones, Ed., Environmental Effects on Engineered Materials, Marcel-Dekker, Inc., 2001 R.H. Jones, Ed., New Techniques for Characterizing Corrosion and Stress Corrosion, TMS, Warrendale, PA, 1995 R.H. Jones, Ed., Stress-Corrosion Cracking Materials Performance and Evaluation, ASM International, 1992 R.D. Kane, Ed., “Environmentally Assisted Cracking: Predictive Methods for Risk Assessment and Evaluation of Materials, Equipment and Structures,” STP 1401, ASTM, 2000

Analysis of SCC •

•

S.M. Bruemmer and L.E. Thomas, Insights into Environmental Degradation Mechanisms from High Resolution Characterization of Crack Tips, Proceedings of Chemistry and Electrochemistry of Corrosion and Stress Corrosion Cracking, R.H. Jones, Ed., TMS, Warrendale, PA, 2001, p 123 R.H. Jones, H. Li, and J.P. Hirth, “Effects of Mixed Mode I/III Loading on Environmental- Induced Cracking,” Corrosion, Vol 57 (No. 1), 2001, p 52

•

•

T. Magnin, D. Tanguy, and D. Delafosse, Physical and Numerical Modeling of the Stress Corrosion Cracking Behavior of Austenitic Stainless Steel, Al-Mg and Nickel- Base Alloys in PWR and Chloride Solutions, Proceedings of Chemistry and Electrochemistry of Corrosion and Stress Corrosion Cracking, R.H. Jones, Ed., TMS, Warrendale, PA, 2001, p 45 R.W. Staehle, Bases for Predicting the Earliest Failures Due to Stress Corrosion Cracking, Proceedings of Chemistry and Electrochemistry of Corrosion and Stress Corrosion Cracking, R.H. Jones, Ed., TMS, Warrendale, PA, 2001, p K-1

Specific Alloys and SCC • • •

N.J.H. Holroyd, Environment-Induced Cracking of High-Strength Aluminum Alloys, EnvironmentInduced Cracking of Metals, R.P. Gangloff and M.B. Ives, Ed., NACE, 1988, p 311 R.H. Jones, D.R. Baer, M.J. Danielson, and J.S. Vetrano, Role of Mg in the Stress Corrosion Cracking of an Al-Mg Alloy, Metals and Mater. Trans A, Vol 32A, 2001, p 1699 N. Sridhar, D.S. Dunn, and A. Anderko, “Prediction of Conditions Leading to Stress Corrosion Cracking of Gas Transmission Lines,” Environmentally Assisted Cracking: Predictive Methods for Risk Assessment and Evaluation of Materials, Equipment and Structures, STP 1401, R.D. Kane, Ed., ASTM, 2000

B. Craig, Hydrogen Damage, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 367–380

Hydrogen Damage Bruce Craig, MetCorr

Introduction HYDROGEN DAMAGE is a form of environmentally assisted failure that results most often from the combined action of hydrogen and residual or applied tensile stress. Hydrogen damage to specific alloys or groups of alloys manifests itself in many ways, such as cracking, blistering, hydride formation, and loss in tensile ductility. For many years, these failures have been collectively termed hydrogen embrittlement; this term persists even though it is improperly used to describe a multitude of failure modes involving hydrogen, several of which do not demonstrate the classical features of embrittlement (that is, reduced load-carrying capability or fracture below the yield strength). This section classifies the various forms of hydrogen damage, summarizes the various theories that seek to explain hydrogen damage, and reviews hydrogen degradation in specific ferrous and nonferrous alloys. Information on the effect of hydrogen on fracture characteristics is available in the article “Modes of Fracture” in Fractography, Volume 12 of ASM Handbook, formerly Metals Handbook, 9th ed.

B. Craig, Hydrogen Damage, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 367–380 Hydrogen Damage Bruce Craig, MetCorr

Classification of Hydrogen Processes The specific types of hydrogen damage have been categorized in order to enhance the understanding of the factors that affect this behavior in alloys and to provide a basis for development and analysis of theories regarding different hydrogen damage mechanisms (Ref 1). Table 1 presents one of these classifications schemes, describing the materials that are susceptible to the various forms of damage, the source of hydrogen, typical conditions for the occurrence of failure, and the initiation site. The mechanism for each of these failure modes is also described briefly. The first three classes are grouped together and designated hydrogen embrittlement because these are the failure modes that typically exemplify classical hydrogen embrittlement.

Table 1 Classifications of processes of hydrogen degradation of metals

Typical materials

Usual source of hydrogen (not exclusive)

Typical conditions

Hydrogen embrittlement Hydrogen Hydrogen stress environment embrittlement cracking Steels, nickel- Carbon and base alloys, low-alloy metastable steels stainless steel, titanium alloys Gaseous H2 Thermal processing, electrolysis, corrosion

Hydrogen attack

Blistering

Shatter cracks, flakes, fisheyes Steels (forgings and castings)

Microperforation

Steels, nickelCarbon and lowbase alloys, alloy steels Be-Cu bronze, aluminum alloys Gaseous Gaseous hydrogen, internal hydrogen from electrochemical charging

Steels, copper, aluminum

Hydrogen sulfide corrosion, electrolytic charging, gaseous

Water vapor reacting with molten steel

Gaseous hydrogen

Loss in tensile ductility

Degradation Metal in flow hydride properties formation

Iron, steels, V, Nb, Ta, Steels (compressors) nickel-base Ti, Zr, U alloys

10-12 to 102 MPa (10-10 to 104 psi) gas pressure

0.1 to 10 ppm total hydrogen content

0.1 to 10 ppm total hydrogen content range of gas pressure exposure

Up to 102 MPa (15 ksi) at 200– 595 °C (400– 1100 °F)

Hydrogen activity equivalent to 0.2 to 1 × 102 MPa (3–15 ksi) at 0–150 °C (30– 300 °F)

Precipitation of dissolved ingot cooling

2 to 8 × 106 MPa (30–125 ksi) at 20– 100 °C (70– 210 °F)

Observed at 100 to 700 °C (-150 to 1290 °F); most severe near 20 °C (70 °F) Strain rate

Observed at 100 to 100 °C (-150 to 210 °F); most severe near 20 °C (70 °F) Strain rate

Observed at 100 to 700 °C (-150 to 1290 °F)

…

…

…

…

Gaseous or Internal internal hydrogen hydrogen from melt; corrosion, electrolytic charging, welding 1–10 ppm 0.1 to 102 hydrogen MPa (15– content 15,000 psi) (iron at 20 gas pressure °C, or 70 hydrogen °F) up to activity 102 MPa (15 must exceed ksi) gaseous solubility hydrogen limit near (various 20 °C (70 metals, T > °F) 0.5 melting point) … …

Occurs in

…

…

…

…

…

…

important; embrittlement more severe at low strain rate; generally more severe in notched or precracked specimens Surface or Failure internal initiation initiation; incubation period not observed Mechanisms Surface or subsurface processes

T, temperature. Source: Ref 1

important; embrittlement more severe at low strain rate; always more severe in notched or precracked specimens Internal crack initiation

absence of effect on yield stress; strain rate important

Internal diffusion to stress concentration

Surface or subsurface processes

Surface and/or internal effect

Surface (decarburization); internal carbide interfaces (methane bubble formation) Carbon diffusion (decarburization); hydrogen diffusion; nucleation and growth (bubble formation)

Internal defect

Internal defect

Unknown

…

Internal defect

Hydrogen diffusion; nucleation and growth of bubble; steam formation

Hydrogen diffusion to voids

Unknown

Adsorption Hydride to precipitation dislocations; solidsolution effects

Hydrogen environment embrittlement occurs during the plastic deformation of alloys in contact with hydrogenbearing gases or a corrosion reaction and is therefore strain-rate dependent. The degradation of the mechanical properties of ferritic steels, nickel-base alloys, titanium alloys, and metastable austenitic stainless steels is greatest when the strain rate is low and the hydrogen pressure and purity are high. Hydrogen stress cracking, also referred to as hydrogen-induced cracking or static fatigue, is characterized by the brittle fracture of a normally ductile alloy under sustained load in the presence of hydrogen. Most often, fracture occurs at sustained loads below the yield strength of the material. This cracking mechanism depends on the hydrogen fugacity, strength level of the material, heat treatment/microstructure, applied stress, and temperature. For many steels, a threshold stress exists below which hydrogen stress cracking does not occur. This threshold is a function of the strength level of the steel and the specific hydrogen-bearing environment. Therefore, threshold stress or stress intensity for hydrogen stress cracking is not considered a material property. Generally, the threshold stress decreases as the yield strength and tensile strength of an alloy increase. Hydrogen stress cracking is associated with absorption of hydrogen and a delayed time to failure (incubation time), during which hydrogen diffuses into regions of high triaxial stress. Hydrogen stress cracking may promote one mode of fracture in an alloy rather than another form normally observed in benign environments. Thus, all modes of cracking have been observed in most commercial alloy systems; however, hydrogen stress cracking usually produces sharp, singular cracks in contrast to the extensive branching observed for stresscorrosion cracking (SCC). The catastrophic cracking of steels in hydrogen sulfide (H2S) environments, referred to as sulfide stress cracking, is a special case of hydrogen stress cracking. Loss in tensile ductility was one of the earliest recognized forms of hydrogen damage. Significant decreases in elongation and reduction in area are observed for steels, stainless steels, nickel-base alloys, aluminum alloys, and titanium alloys exposed to hydrogen. This mode of failure is most often observed in lower-strength alloys, and the extent of loss in tensile ductility is a function of hydrogen content of the material. Loss in tensile ductility behavior is strain-rate sensitive and becomes more pronounced as the strain-rate decreases. Hydrogen attack is a high-temperature form of hydrogen damage that occurs in carbon and low-alloy steels exposed to high-pressure hydrogen at high temperatures for extended time. Hydrogen enters the steel and reacts with carbon either in solution or as carbides to form methane gas; this may result in the formation of cracks and fissures or may simply decarburize the steel, resulting in a loss in strength of the alloy. This form of damage is temperature dependent, with a threshold temperature of approximately 200 °C (400 °F). Blistering occurs predominantly in low- strength alloys when atomic hydrogen diffuses to internal defects, such as laminations or nonmetallic inclusions, and then precipitates as molecular hydrogen (H 2). The pressure of molecular hydrogen can attain such high values that localized plastic deformation of the alloy occurs, forming a blister that often ruptures. Blisters are frequently found in low-strength steels that have been exposed to aggressive corrosive environments (such as H2S) or cleaned by pickling. Shatter cracks, flakes, and fisheyes are features common to hydrogen damage in forgings, weldments, and castings. They are attributed to hydrogen pickup during melting operations when the melt has a higher solubility for hydrogen than the solid alloy. During cooling from the melt, hydrogen diffuses to and precipitates in voids and discontinuities, producing the features that result from the decreased solubility of hydrogen in the solid metal. In many aspects, these features are comparable to blistering, and this could be considered a special case of that class. Microperforation by high-pressure hydrogen occurs at extremely high pressures of hydrogen near room temperature. Microperforation occurs predominately in steels. This form of hydrogen damage manifests itself as a network of small fissures that allows permeation of the alloy by gases and liquids. Degradation in flow properties in hydrogen environments has been found at ambient temperatures for iron and steel and at elevated temperature for several alloy systems. The steady- state creep rate under constant load has been observed to increase in the presence of hydrogen for some nickel-base alloys. Hydride formation produces embrittlement in magnesium, tantalum, niobium, vanadium, uranium, thorium, zirconium, titanium, and their alloys, as well as many other less common metals and alloys. The degradation of mechanical properties and the cracking of these metals and their alloys are attributable to the precipitation of metal hydride phases. Hydrogen pickup often results from welding, heat treating, charging from corrosion processes, or during melting of the alloy. Hydride formation is enhanced for some metal-hydrogen systems by the application of stress—the so-called stress-induced hydride formation. Alloy systems that form hydrides are generally ductile at high, >300 K (80 °F), and low, 0.22 … Volume … … Depends on … exothermicity of the dissolution of H by the particle i i (a) ε H is the interaction coefficient. A negative ε H means hydrogen is attracted. (b) Values of interaction energies are either experimental or are calculated (when between parentheses) at room temperature. Source: Ref 13

References cited in this section 2. C. Zapffe and C. Sims, Hydrogen Embrittlement, Internal Stress and Defects in Steel, Trans. AIME, Vol 145, 1941, p 225

3. N.J. Petch and P. Stables, Delayed Fracture of Metals Under Static Load, Nature, Vol 169, 1952, p 842 4. A.R. Troiano, The Role of Hydrogen and Other Interstitials in the Mechanical Behavior of Metals, Trans. ASM, Vol 52, 1960, p 57 5. R.A. Oriani, A Mechanistic Theory of Hydrogen Embrittlement of Steels, Bunsen-Gesellschaft Phys. Chemie, Vol 76, 1972, p 848 6. J.F. Knott, Fracture Toughness and Hydrogen Assisted Crack Growth in Engineering Alloys, Hydrogen Effects in Materials, A.W. Thompson and N.R. Moody, Ed., TMS, 1996, p 387 7. C.D. Beachem, A New Model for Hydrogen Assisted Cracking (Hydrogen Embrittlement), Metall. Trans., Vol 3, 1972, p 437 8. J.P. Hirth, The Role of Hydrogen in Enhancing Plastic Instability and Degrading Fracture Toughness of Steels, Hydrogen Effects in Materials, A.W. Thompson and N.R. Moody, Ed., TMS, 1996, p 507 9. S. Gahr, M.L. Grossbech, and H.K. Birnbaum, Acta Metall., Vol 25, 1977, p 125 10. F.H. Vitovec, Modeling of Hydrogen Attack of Steel in Relation to Material and Environmental Variables, Current Solutions to Hydrogen Problems in Steels, C.G. Interrante and G.M. Pressouyre, Ed., American Society for Metals, 1982 11. G.M. Pressouyre and I.M. Bernstein, A Quantitative Analysis of Hydrogen Trapping, Metall. Trans. A, Vol 9, 1978, p 1571 12. W.W. Gerberich, Effect of Hydrogen on High Strength and Martensitic Steels, Hydrogen in Metals, I.M. Bernstein and A.W. Thompson, Ed., American Society for Metals, 1974, p 115 13. G.M. Pressouyre, A Classification of Hydrogen Traps in Steel, Metall. Trans. A, Vol 10, 1979, p 1571


Hydrogen Damage in Iron-Base Alloys Pure Irons. Hydrogen damage may occur in relatively pure irons, such as Ferrovac E and other magnetically soft irons, producing either transgranular or intergranular fracture, depending on the heat treatment and the presence of impurities and solutes (carbon, oxygen, and nitrogen) at the grain boundaries. Hydrogen can also lower the yield strength and flow stress of high-purity iron at approximately room temperature or lead to hardening. Although the yield strength is reduced for impure iron, the flow stress may be increased under conditions of low hydrogen fugacity. Ferrous Alloys. Many factors affect the behavior of ferrous alloys in hydrogen-bearing environments. Hydrogen concentration, temperature, heat treatment/microstructure, stress level (applied and yield stress), solution composition, and environment are the primary factors involved in determining susceptibility to hydrogen embrittlement. Figure 2 shows the effect of hydrogen concentration on the time to failure for a high-

strength steel. The longer the baking time, the lower the residual hydrogen in the steel matrix. In general, increasing the concentration of hydrogen in an alloy reduces time to failure and the stress level at which failure occurs.

Fig. 2 Static fatigue curves for various hydrogen concentrations obtained by different baking times at 150 °C (300 °F). Sharp-notch high-strength steel specimens 1590 MPa (230 ksi); normal notch strength: 2070 MPa (300 ksi). Source: Ref 4 Hydrogen concentration in the alloy is a function of the fugacity or the approximate concentration of hydrogen at the surface exposed to the environment. Therefore, hydrogen embrittlement is controlled by the hydrogen gas pressure or pH of the environment as well as constituents within the environment that may accelerate or inhibit the entry of hydrogen into the alloy. Elements such as sulfur, phosphorus, antimony, tin, and arsenic and their compounds have been found to inhibit the hydrogen recombination reaction in aqueous solutions, thus increasing the charging of atomic hydrogen into the alloy. In contrast, small amounts of oxygen in gaseous hydrogen environments have demonstrated an inhibitive effect on crack growth of high- strength steels subject to hydrogen cracking. Figure 3 shows the dependence of both the threshold stress intensity and the crack growth rate of high-strength American Iron and Steel Institute (AISI) 4130 (Unified Numbering System, or UNS, G41300) steel on hydrogen pressure. Increasing the hydrogen pressure reduces the threshold stress intensity for crack initiation and increases the crack growth rate for a specific stress-intensity value.

Fig. 3 The stress-intensity K dependence of crack growth rate da/dt at various hydrogen pressures at 24 °C (75 °F) for AISI 4340 steel. Source: Ref 14 Temperature also plays an important role in the hydrogen embrittlement of ferrous alloys. Embrittlement is most severe near room temperature (Ref 15) and becomes less severe or nonexistent at higher or lower temperatures (Fig. 4). At lower temperatures, the diffusivity of hydrogen is too sluggish to fill sufficient traps, but at high temperatures, hydrogen mobility is enhanced, and trapping is diminished. As can be seen in Fig. 4, embrittlement is also strongly strain-rate dependent. At high strain rates, fracture may proceed without the assistance of hydrogen, because the mobility of hydrogen is not sufficient to maintain a hydrogen atmosphere around moving dislocations.

Fig. 4 Notch tensile strength of high-strength steel plotted against testing temperature for three strain rates (crosshead speeds, ). Source: Ref 15 Figure 5 shows threshold stress intensity as a function of yield strength for a large number of steels in a variety of hydrogen-containing environments. It can be seen that the threshold stress intensity for crack growth generally decreases with increasing yield strength, regardless of environment, and that very high-strength steels are not usable in hydrogen environments. Threshold stress intensities and crack growth rates are a function of the specific hydrogen environments, with H2S being one of the most severe environments (Fig. 6). At lower yield strengths, the mechanism for hydrogen-assisted failure apparently changes, and blistering becomes the more common feature of failure. The threshold stress intensities for high-strength steels subjected to hydrogen environments are significantly less than those thresholds measured under benign conditions. These lower thresholds lead to subcritical crack growth when compared to critical values expected from fracture mechanics. Therefore, it is common to designate these thresholds as KISCC or KIH.

Fig. 5 Lower bounds on threshold stress intensity versus yield strength for ferrite-pearlite, bainitic, and martensitic steels stressed in five hydrogen-producing environments. JSW, Japan Steel Works. Source: Ref 16

Fig. 6 Crack velocity as a function of stress intensity for a chromium-molybdenum-vanadium steel at 291 K (18 °C, or 64 °F). Source: Ref 17 In low-strength steels, 700 MPa (100 ksi) yield strength or less, hydrogen damage occurs predominately by loss in tensile ductility or blistering. For loss in tensile ductility, hydrogen promotes the formation and/or growth of voids by enhancing the decohesion of the matrix at carbide particle and inclusion interfaces. At higher hydrogen fugacities and often in the absence of stress, blistering or a form of cracking also associated with inclusions—referred to as stepwise cracking, blister cracking, or hydrogen- induced cracking (HIC)—can occur. Stepwise cracking has been observed frequently in low- strength steels subjected to H2S-containing environments in the absence of stress (Fig. 7). This type of cracking is not dependent on steel strength but is strictly a function of steel chemistry, processing, and the severity of the hydrogen environment (Ref 18). In H2S environments, a cracking morphology has been observed in low-strength steels that combines the features of

hydrogen stress cracking and HIC and has been designated stress-oriented hydrogen-induced cracking (SOHIC) (Ref 19). It is unfortunate that too often when investigators observe a new morphology of cracking, they establish a new name for the cracking regardless whether this is justified mechanistically. The result is that the literature is cluttered with many names for the same phenomenon.

Fig. 7 Stepwise cracking of a low-strength pipeline steel exposed to H2S. 6× Fracture of low-strength steels in hydrogen environments may be characterized by ductile dimple rupture, tearing, cleavage, quasi-cleavage, and, less frequently under certain conditions, intergranular cracking. The article “Modes of Fracture” in Fractography, Volume 12 of ASM Handbook, formerly Metals Handbook, 9th ed., contains fractographs illustrating the effects of hydrogen on the fracture appearance of steels. High-strength steels, those with yield strengths above 700 MPa (100 ksi), are prone to fracture either in an intergranular fashion or by quasi-cleavage, depending on the stress intensity. These steels commonly display an incubation time before fracture initiates under sustained loading, usually in association with regions of high triaxial stress. Because triaxial stresses are created at notch roots or under plane strain, fracture initiates internally in the steel. Intergranular fracture is promoted by the presence of impurity elements at prior-austenite or -ferrite grain boundaries. Elements such as phosphorus, sulfur, tin, antimony, and arsenic have been found to enhance the intergranular fracture of high- strength steels in hydrogen, and as expected, temper-embrittled steels are even more susceptible to hydrogen stress cracking than steels that are not embrittled. Metallurgical structure can have a profound effect on the resistance of steels to hydrogen embrittlement. When compared at equivalent strength levels, a quenched-and-tempered fine- grain microstructure is more resistant to cracking than a normalized or bainitic steel. However, this is also dependent to a large extent on the strength level at which this comparison is made. In general, the most resistant microstructure is a highly tempered martensitic structure with equiaxed ferrite grains and spheroidized carbides evenly distributed throughout the matrix. Because microstructure is dependent on heat treatment and composition, these factors are not easily separated and must be considered together. There is also a grain size effect that produces enhanced resistance to hydrogen with decreasing prior- austenite grain size (Fig. 8). However, if the grain size is significantly larger than the plastic zone size, the reverse may be true.

Fig. 8 Illustration of how a refinement in grain size improves resistance to hydrogen failure as measured by the time to failure of two strengths of AISI 4340 steels. Source: Ref 20 The role of alloying elements is quite complex and not easily distinguished from the effects of heat treatment, microstructure, and strength level. Depending on the microstructure and strength level, a specific alloying element may or may not contribute to the hydrogen embrittlement resistance of an alloy or may even increase susceptibility to cracking. The concentration of the alloying element is also a factor in the behavior of alloys in hydrogen. Molybdenum, for example, is beneficial in reducing susceptibility to sulfide stress cracking, a form of hydrogen stress cracking, in AISI 4130 steels up to approximately 0.75 wt%. However, beyond this concentration, a separate Mo2C phase precipitates in the alloy after tempering at above 500 °C (930 °F), significantly reducing resistance to sulfide stress cracking. In general, elements such as carbon, phosphorus, sulfur, manganese, and chromium impart greater susceptibility to hydrogen embrittlement in low-alloy steels. However, large increases in such elements as chromium, nickel, and molybdenum in order to produce stainless steels alter the crystal structure, microstructure, and subsequently the heat treatment requirements and therefore the hydrogen embrittlement behavior of this group of ferrous alloys. Stainless Steels The response of stainless steels to hydrogen-bearing environments is basically related to their strength level. Ferritic stainless steels have excellent resistance to hydrogen embrittlement because of their low strength and enhanced ductility. However, if the ferritic stainless steels are cold worked, they may become susceptible to cracking in hydrogen environments. Similarly, austenitic stainless steels are highly resistant to hydrogen cracking in the annealed or lightly cold-worked condition but can become quite susceptible when heavily cold worked. This increased susceptibility to hydrogen cracking due to increasing yield strength from cold working is similar to the dependence of carbon and low-alloy steel on strength. Decreased resistance to hydrogen for highly cold- worked austenitic stainless steels is largely attributed to the deformation-induced formation of martensite. For those austenitic stainless steels having a very stable austenite phase and high yield strength, such as 21Cr-6Ni-9Mn (UNS S21900), susceptibility is considered to be solely a function of yield strength, similar to the body- centered cubic (bcc) low-alloy steel behavior. Other factors that may affect the susceptibility of austenitic stainless steels to hydrogen damage are the possible formation of metastable hydride phase that would produce a hydride-based fracture path and the interaction of

hydrogen with stacking faults to reduce stacking fault energy in the austenite, leading to planar slip and brittle fracture. The degree of participation of any of these factors has not been established. Just as a similarity exists between austenitic stainless steels and low-alloy steels at the high- strength end of the spectrum, the lower-strength austenitics behave in the same manner as the low-alloy steels in hydrogen by a reduction in ductility. Figure 9 shows the loss in reduction in area for several austenitic stainless steels in highpressure hydrogen. It is apparent that a wide variation in hydrogen damage exists between these various austenitic alloys. Type 304L (UNS S30403) is the most susceptible to loss in tensile ductility, and the stable austenitic alloys, such as 15Cr-25Ni (UNS S66286), are almost unaffected. As observed in carbon and lowalloy steels, there is a temperature effect (~0 °C, or 32 °F) involved with the ductility loss in austenitic stainless steels, although it is somewhat lower than the room-temperature dependence observed for low-alloy steels.

Fig. 9 Ductility loss for several austenitic stainless steels in high-pressure hydrogen. Source: Ref 21 Duplex stainless steels represent an interesting type of behavior in hydrogen as a result of their two-phase structure of ferrite and austenite. Because the solubilities and diffusivities of hydrogen are quite different between the austenite and ferrite phases, the response of the overall duplex alloy is significantly different than either a ferritic or austenitic stainless steel. Reference 22 compiles much of the data in the literature regarding the measured hydrogen diffusion coefficient in duplex stainless steels, as shown in Fig. 10. As could be expected, there is a large degree of scatter in these data as a result of factors such as the relative ferrite-toaustenite ratio, the grain- boundary areas, the volume fraction of precipitates, and so on.

Fig. 10 Published values of hydrogen diffusion coefficients in duplex stainless steels. Data sources are identified in Ref 22. Curves 4 and 5 are from the same source. Curves 8 and 9 are from the same source: Ref 22 A theoretical analysis for hydrogen transport in these two-phase alloys (Ref 23) showed that the effective diffusivity of hydrogen in duplex stainless steels was reduced by a factor of approximately 400 compared to a fully ferritic steel, and that trapping at the ferrite-austenite interface was the most significant factor in explaining this reduced diffusivity of hydrogen in duplex stainless steels relative to ferritic steels. The behavior of hydrogen in duplex stainless steels is very complex, because a change in the ferrite- to-austenite ratio not only alters the relative interface area but also changes the volume fraction of each phase, thereby increasing the solubility and diffusivity of hydrogen in the alloy as a whole. It is apparent that there are many similarities in behavior between stainless steels and carbon and low-alloy steels in hydrogen. Similarly, the martensitic and precipitation-hardening stainless steels exhibit the same dependence on strength level and microstructure as observed in low-alloy steels. Martensitic and precipitationhardening stainless steels are extremely susceptible to hydrogen embrittlement with increasing yield strength. Figure 11 compares several grades of precipitation-hardening stainless steel with type 410 martensitic stainless steel tested in an aqueous environment saturated with H2S. The numbers adjacent to each data point represent the tempering or aging treatment. Generally, the same trend of decreasing time to failure with increasing yield strength is observed as for low- alloy steels. The poor performance of the type 410 martensitic stainless steel

compared to precipitation-hardening stainless steels is typical of the behavior of most martensitic stainless steels, which also compare poorly with low-alloy steels at the same strength level. Even the newer family of martensitic stainless steels, recently introduced as Super 13 Cr, have relatively poor resistance to cracking in hydrogen-bearing environments. Figure 12 shows the regions of hydrogen-assisted cracking (HAC) and those that are not (no HAC) for a steel containing approximately 15% Cr, 4% Ni, and 1.5% Mo. Note that the hydrogen-generating power of the solution, given by the combination of pH and H2S concentration, defines the boundary of cracking (Ref 25).

Fig. 11 Time to failure of various alloys as a function of yield strength when tested under 345 MPa (50 ksi) of applied stress in saturated H2S. Numbers adjacent to data points represent tempering or aging treatment; parenthetical values indicate approximate amounts of austenite. Source: Ref 24

Fig. 12 The H2S-pH tolerance of low-carbon martensitic stainless steel tested by the slow strain-rate technique. HAC, hydrogen-assisted cracking. Source: Ref 25 Ultrahigh-strength, >1400 MPa (200 ksi), martensitic stainless steels, low-alloy steels, and maraging steels are extremely susceptible to cracking in hydrogen environments, including aqueous solutions containing NaCl. Although chlorides are the primary cause of SCC in many alloy systems, it is generally accepted that the mechanism of cracking in ultrahigh-strength steels is related to hydrogen embrittlement. Figure 13 compares the crack growth rate of AISI 4340 in several environments against grade 250 maraging steel. The maraging steel has much better resistance to crack propagation in 3.5% NaCl than 4340 and displays a high threshold stress intensity for crack propagation. For maraging steels, there is evidence that peak-aged condition or slight overaging may improve resistance to hydrogen embrittlement, but underaging is detrimental to resistance.

Fig. 13 Effect of material, environment, and stress-intensity level on crack growth. Source: Ref 12 Hydrogen attack is strictly a high-temperature form of hydrogen damage that primarily affects ferritic steels. Hydrogen attack is dependent on time, temperature, and hydrogen pressure. After a certain incubation time, a diminution of properties occurs with the onset of decarburization and cracking (Fig. 14). Resistance of steels to hydrogen attack is related to the stability of the carbides, and as such, the addition of carbide- stabilizing elements such as vanadium, titanium, niobium, and molybdenum enhance the resistance to this form of hydrogen damage.

Fig. 14 Relative change of properties of a Society of Automotive Engineers (SAE) 1020 steel as a function of time of exposure to hydrogen at 427 °C (800 °F) and 6.2 MPa (900 psi) partial pressure. Source: Ref 10 Chromium-molybdenum steels have been found to be the most resistant to hydrogen attack for the cost involved; therefore, a great deal of industrial experience has been gained with these alloys. Much of this experience has been compiled and plotted to provide a guideline for steel selection as a function of temperature and hydrogen partial pressure and is presented as a series of curves often referred to as a Nelson diagram (Ref 26). As more experience is gathered on the use of these steels in hydrogen, this series of curves is periodically updated by the American Petroleum Institute.

References cited in this section 4. A.R. Troiano, The Role of Hydrogen and Other Interstitials in the Mechanical Behavior of Metals, Trans. ASM, Vol 52, 1960, p 57 10. F.H. Vitovec, Modeling of Hydrogen Attack of Steel in Relation to Material and Environmental Variables, Current Solutions to Hydrogen Problems in Steels, C.G. Interrante and G.M. Pressouyre, Ed., American Society for Metals, 1982 12. W.W. Gerberich, Effect of Hydrogen on High Strength and Martensitic Steels, Hydrogen in Metals, I.M. Bernstein and A.W. Thompson, Ed., American Society for Metals, 1974, p 115 14. H.G. Nelson and D.P. Williams, Quantitative Observations of Hydrogen Induced Slow Crack Growth in a Low Alloy Steel, Stress Corrosion Cracking and Hydrogen Embrittlement of Iron Base Alloys, National Association of Corrosion Engineers, 1977, p 390 15. B.A. Graville, R.G. Baker, and F. Watkinson, Br. Weld. J., Vol 14, 1967, p 337 16. R.P. Gangloff, A Review and Analysis of the Threshold for Hydrogen Environment Embrittlement of Steel, Corrosion Prevention and Control, Proc. 33rd Sagamore Army Materials Research Conf. (MA), U.S. Army Materials Technology Laboratory, 1986 17. P. McIntyre, The Relationship Between Stress Corrosion Cracking and Sub Critical Flaw Growth in Hydrogen and Hydrogen Sulphide Gases, Stress Corrosion Cracking and Hydrogen Embrittlement of Iron Base Alloys, National Association of Corrosion Engineers, 1977, p 788 18. A. Ikeda, “Progress on the Metallurgical Investigation of HIC Phenomena and Development of HIC Resistant Steel,” Second International Conf. on Interaction of Steels with Hydrogen in Petroleum Industry Pressure Vessel and Pipeline Service, 19 Oct 1994 (Vienna)

19. “Guidelines on Materials Requirements for Carbon Steels and Low Alloy Steels for H2S Containing Environments in Oil and Gas Production,” European Federation of Corrosion Publication 16, Institute of Materials, London, 1995 20. R.P.M. Procter and H.W. Paxton, Trans. ASM, Vol 62, 1969, p 989 21. A.W. Thompson, “Hydrogen Embrittlement of Stainless Steels and Carbon Steels,” paper presented at the midyear meeting (Toronto, Canada), American Petroleum Institute, 1978 22. T. Boellinghaus, H. Hoffmeister, and C. Middel, “Scatterbands for Hydrogen Diffusion in Steels Having a Ferritic or Martensitic Microstructure and Steels Having an Austenitic Microstructure at Room Temperature,” Weld. World, Vol 37 (No.1), 1996, p 16–23 23. A. Turnbull and R.B. Hutchings, “Analysis of Hydrogen Atom Transport in a Two- Phase Alloy, Mater. Sci. Eng., A177, 1994, p 161–171 24. R.R. Gaugh, “Sulfide Stress Cracking of Precipitation Hardening Stainless Steels,” Paper 109, presented at Corrosion/77, National Association of Corrosion Engineers, 1977 25. T. Boellinghaus, H. Hoffmeister, and S. Dietrich, “Slow Strain Rate Testing of Low Carbon Martensitic Stainless Steels,” EuroCorr '97, Sept 1997 (Tondheim, Norway), Institute of Materials, London 26. “Steels for Hydrogen Service at Elevated Temperatures and Pressures in Petroleum Refineries and Petrochemical Plants,” Publication 941, 2nd ed., American Petroleum Institute, June 1977


Nickel Alloys Nickel and its alloys are susceptible to hydrogen damage in both aqueous and gaseous hydrogen environments. The same factors that affect hydrogen embrittlement susceptibility in ferrous alloys are also prevalent in nickel alloys, although to a lesser degree. In general, fcc metals, because of their greater ease of slip and reduced solute diffusivities as compared to bcc materials, are less susceptible to hydrogen damage. As with ferrous alloys, hydrogen in nickel and its alloys may introduce intergranular, transgranular, or quasi-cleavage cracking, and although the macroscopic features appear to be brittle, on a microscopic scale there is a high degree of local plasticity, suggesting that hydrogen enhances flow at the crack tip. Figure 15 shows the reduction in ductility associated with hydrogen charging a 72Ni-28Fe alloy as a function of strain rate. Almost identical behavior was observed for unalloyed nickel.

Fig. 15 Ductility at fracture as a function of strain rate in a hydrogen-charged and uncharged 72Ni-28Fe alloy. RA, reduction in area. Source: Ref 4 Alloys based on the ternary Fe-Ni-Cr system (Incoloy) and Inconel alloys show reductions in ductility when charged with hydrogen, depending on the specific thermomechanical treatments performed on the alloy. Generally, these alloys are more resistant to hydrogen stress cracking in the cold-worked and unaged condition as compared to the cold-worked and aged condition. Age-hardenable alloys show the least resistance to hydrogen when aged to their peak or near- peak strength. Stabilization treatments followed by aging reduce the resistance of these alloys as compared to direct aging. Increased cold work also produces a loss in tensile ductility. Hydrogen stress cracking of these alloys may also occur, rather than a loss in tensile ductility, when they possess high yield strengths or are under high hydrogen fugacity. The nickel-copper alloys (Monels) have also been found to be susceptible to hydrogen embrittlement, with increasing strength obtained by cold working or aging. Figure 16 is a good comparison of the effect of hydrogen on the threshold cracking behavior of several stainless steels and nickel alloys (Ref 27). It is interesting to note that there is not an obvious distinction between singlephase alloy resistance and that of multiphase alloys. For example, the single-phase alloy 625 performs much the same as the precipitation-hardening alloy 718. Likewise, the single-phase alloy 310 is not significantly different from the age-hardening alloy A286. The significant reductions in the threshold stress-intensity factor (Kth) for types 301 and 304 are attributable to the strain-induced transformation to martensite in these alloys.

Fig. 16 Effect of hydrogen on the threshold stress intensity for crack propagation, normalized against the hydrogen-free values. Ref 27 Several nickel-base alloys are resistant to hydrogen stress cracking when cold worked to yield strengths in excess of 1240 MPa (180 ksi). However, when these alloys (Hastelloy alloy C- 276, Hastelloy alloy C-4, and Inconel alloy 625) are aged at low temperature, their resistance to hydrogen cracking is considerably diminished. This behavior is attributed to the segregation of phosphorus and sulfur to grain boundaries, which provide low-energy fracture paths (much the same as occurs in high-strength steels), or to an ordering reaction of the form Ni2(Cr, Mo). Figure 17 presents data that relate the reduction in area loss to ordering for the ordering alloy Ni2Cr charged with hydrogen.

Fig. 17 Effect of the degree of order on the embrittlement susceptibility of Ni2Cr. Regions of intergranular (IG) and ductile transgranular fracture (TG) are shown. RA, reduction of area. Source: Ref 28


4. A.R. Troiano, The Role of Hydrogen and Other Interstitials in the Mechanical Behavior of Metals, Trans. ASM, Vol 52, 1960, p 57 27. P.D. Hicks and C.J. Altstetter, “Hydrogen- Enhanced Cracking of Superalloys,” Metall. Trans. A, Vol 32, 1992, p 237–248 28. B.J. Berkowitz, M. Kurkela, and R.M. Latanision, Effect of Ordering on the Hydrogen Permeation and Embrittlement of Ni2Cr, Hydrogen Effects in Metals, I.M. Bernstein and A.W. Thompson, Ed., American Society for Metals, 1981, p 411


Aluminum Alloys Only recently has it been determined that hydrogen embrittles aluminum. For many years, all environmental cracking of aluminum and its alloys was represented as SCC; however, testing in specific hydrogen environments has revealed the susceptibility of aluminum to hydrogen damage. Hydrogen damage in aluminum alloys may take the form of intergranular or transgranular cracking or blistering. Blistering is most often associated with the melting or heat treatment of aluminum, where reaction with water vapor produces hydrogen. Blistering due to hydrogen is frequently associated with grain-boundary precipitates or the formation of small voids. Blister formation in aluminum is different from that in ferrous alloys in that it is more common to form a multitude of near-surface voids that coalesce to produce a large blister. In a manner similar to the mechanism in iron- base alloys, hydrogen diffuses into the aluminum lattice and collects at internal defects. This occurs most frequently during annealing or solution treating in air furnaces prior to age hardening. Dry hydrogen gas is not detrimental to aluminum alloys; however, with the addition of water vapor, subcritical crack growth increases dramatically (Fig. 18). The threshold stress intensity for cracking of aluminum also decreases significantly in the presence of humid hydrogen gas at ambient temperature (Fig. 19).

Fig. 18 Effect of humidity on subcritical crack growth of high-strength aluminum alloys in hydrogen gas. Source: Ref 29

Fig. 19 Crack velocity of four high-strength aluminum alloys plotted as a function of crack-tip stress intensity in moist and dry hydrogen gas. Source: Ref 29 Crack growth in aluminum in hydrogen is also a function of hydrogen permeability, as in the iron-and nickelbase alloys. Hydrogen permeation and the crack growth rate are a function of potential, increasing with more negative potentials, as expected for hydrogen embrittlement behavior. Similarly, the ductility of aluminum alloys in hydrogen is temperature dependent, displaying a minimum in reduction in area below 0 °C (32 °F); this is similar to other fcc alloys (Ref 29). Most of the work on hydrogen embrittlement of aluminum alloys has been on the 7000 series (Al-Zn-Mg); therefore, the full extent of hydrogen damage in aluminum alloys has not been determined or the mechanisms established. Some evidence for a metastable aluminum hydride has been found that would explain the brittle intergranular fracture of aluminum-zinc-magnesium alloys in water vapor. However, the instability of the hydride is such that is has been difficult to evaluate. Another explanation for intergranular fracture of these alloys is preferential decohesion of grain boundaries containing segregated magnesium. Overaging of these alloys increases their resistance to hydrogen embrittlement in much the same way as for highly tempered martensitic steels.

Reference cited in this section

29. M.O. Speidel, Hydrogen Embrittlement of Aluminum Alloys, Hydrogen in Metals, I.M. Bernstein and A.W. Thompson, Ed., American Society for Metals, 1974


Copper Alloys A common form of hydrogen damage in copper has been known for years as steam embrittlement and is observed only when copper contains oxygen. Hydrogen entering the metal reacts with oxygen either in solid solution or at oxide inclusions to form water. At temperatures above the critical temperature for water, steam forms, and the pressure generated is sufficient to produce microcavity formation and cracking. The equation for reaction with cuprous oxide particles is Cu2O + 2H = 2Cu + H2O(g). The circumstances under which this form of hydrogen damage occurs are typically related to the annealing of copper in a hydrogen atmosphere. At lower temperatures, steam is not generated and therefore does not create the problem. Although the use of oxygenfree copper essentially eliminates susceptibility to steam embrittlement, the heavy cold working of oxygen-free copper can result in grain-boundary void formation when hydrogen is introduced. Compared to the ferrous and nickel alloys, relatively little work has been performed specific to the hydrogen embrittlement of copper-base alloys. There is some indication that age-hardenable alloys such as berylliumcopper are susceptible to embrittlement under severe charging conditions, producing a loss in tensile ductility. A systematic study of these alloys remains to be performed.


Titanium Alloys Titanium and its alloys suffer hydrogen damage primarily by hydride-phase formation. Pure α-titanium is relatively unaffected by small concentrations (180 1.5 High (8.9) 30 86 >180 1.3 At the upper end of the cooling water pH range (9.0), silica solubility increases to 115 ppm (20 °C, or 68 °F) and 140 ppm (30 °C, or 86 °F). A control limit of 180 ppm corresponds to saturation levels of 1.5 and 1.3, respectively. In systems where concentration ratio is limited by silica solubility, it is recommended that the concentration ratio limit be reestablished seasonally based on amorphous silica saturation level or whenever significant temperature changes occur (Ref 19). pH level

Boundary-Element Modeling Early predictions of corrosion rates and estimates of adequate cathodic protection (CP) have traditionally been based on case studies and sample exposure tests. Applying these techniques to real structures usually involves extrapolations, use of large safety factors, and ongoing corrections and maintenance of the system. In the late 1960s, the finite-element method was applied to the problem by discretization of the electrolytically conductive environment with a mesh (Ref 20). The mathematical solution is found at the intersection points, or nodes, of the mesh. The solution is numerical, because an approximate form of Laplace's equation governs variations of the solution between adjacent nodes. However, creating a finite-element mesh can be an extremely tedious and

time-consuming process. Even when the mesh-generation process is automated, it is difficult to perform a simulation when there are large geometry scale differences in the model, which is exactly the case found in most corrosion control problems where anodes are small compared with the size of the structure and the problem areas are most likely in corners and areas of complex geometry. In the late 1970s, boundary-element (BE) methods became available. As the name implies, this numerical method requires mesh elements to be created but now only on the boundary (or surfaces) of the object to be modeled (Fig. 8). The main advantages of boundary elements for impressed current cathodic protection (ICCP) analysis are (Ref 20): •

•

Meshes are now only on the surface; hence, only two-dimensional elements are required. Mesh generators can be used with confidence, and models can be constructed extremely quickly and inexpensively once the geometry is defined. Models can be created with fine detail of key or complex areas while course modeling the remaining structure or large volume of electrolyte. Boundary-element methods are very effective and accurate for modeling large or infinite domains, as is the case for CP analysis.

•

Fig. 8 Boundary-element model of the wetted surface of the propeller section of a ship showing the dark anode area in contrast with the protected cathodic surface The design goal of an ICCP system is to produce an evenly distributed protection potential on the structure as well as to reduce the power consumption of the anodes to a minimum. The available design variables are the number of anodes, their location, and the location of the reference cells. The constraints on the design are the values of the potential on the structure. In order to provide adequate protection, the potential must be less than a specified value, for example, -800 mV. In order to prevent overprotection, the potential must be greater than a specified value, for example, -900 mV. By combining an automatic optimization procedure with the BE model of the ICCP system, an optimal solution can be obtained. The equation governing electrochemical corrosion for the wetted surface of a ship hull is (Ref 21): k

2

Φ=0

(Eq 2)

where Φ is the potential, k is the conductivity of the electrolyte, and 2 is the symbol for a Laplacian operator, a second-order, multidimensional, partial differential equation. Equation 2 is valid if the electrolyte is homogeneous, there are no electrical sources or sinks, and the system is electroneutral. A shipboard ICCP system can be modeled in such a way to meet these conditions. Seawater is often represented as a uniform mixture of multiple components. Current source points and exposed metal can be represented by boundary conditions, eliminating the need to include sources and sinks in the model. Electroneutrality maintains charge equilibrium for the ship, surrounding water, and ICCP system. The solution space for the problem defined by Eq 2 used in the BE approach is the surface, Γ, which bounds the domain, Ω, as defined by: Γ = ΓA + ΓC + ΓI

(Eq 3)

where ΓA is the anodic surface, ΓC is the cathodic surface, and ΓI is the insulated surface. The surface, Γ, must be continuous, but all sections of one surface type do not have to be contiguous. An ICCP system consists of the surfaces to be protected, the anodes, the reference cells, and the power supply. Anodes are defined by maintaining the potential at a constant value, ΦA: Φ(x, y) = ΦA

(Eq 4)

defining the current density as a constant, qA, on a surface: (Eq 5) where Φ(x, y) is the electrical potential at the point (x, y), and n(x, y) is the normal to the surface at the point (x, y). Reference cells are defined as specific points on the hull where the mathematical solution is obtained. It has been demonstrated that BE modeling can accurately predict experimental results. A major issue for accurate predictions is the availability of detailed polarization data. However, the lack of availability of accurate polarization data for a particular design condition does not eliminate all advantages associated with BE modeling. Reasonable polarization data can be used to obtain potential maps that identify good and bad regions of protection. Boundary-element methods also can be used to evaluate the effect of a single parameter on system performance. In this way, basic understanding of electrochemical corrosion and parameter interactions can be obtained. Several parametric studies have been completed to date (Ref 21): • • • •

Damage levels in the propeller area Seawater conductivity Finite paint resistance Influence of stray current source on system performance

It is possible to design experiments that would account for the majority, if not all, of the factors that influence polarization response for the materials involved in ICCP. However, this approach quickly leads to impracticably large experimental programs. There are two processes for ICCP system design that rely on BEs but eliminate the need for a high degree of accuracy in polarization response. One proposed method defines the polarization response as an unknown to be solved in the BE process (Ref 21). Sensors provide potential information at specific locations. An inverse problem is defined in which the potential map of the ship hull is defined from the measured data and assumptions of behavior in the regions between sensor points. The solution, obtained through a series of BE evaluations, is the polarization data. Discrete areas of damage can be located based on differences in calculated potentials and measured values at the sensor locations. Once the polarization response is determined for a given ship geometry and service condition, anode strengths and power requirements can be readily obtained. Combined Boundary Element and Physical Scale. The second proposed approach is a hybrid design combining BE and physical scale (PS) modeling techniques (Ref 21). In the first stage, BE modeling is used to develop best estimates of the layout of an ICCP system. Anode locations, anode numbers, reference cell locations, and reference cell numbers are factors that can be varied using computational analyses. The polarization response used in this phase does not have to duplicate service conditions but only has to be a reasonable approximation. The second stage of the ICCP design uses PS modeling. In PS modeling, the structural dimensions and the conductivity of the electrolyte are scaled by the same factor (Ref 22). In this process, the scaled model and the full-sized structure maintain identical current density values at points, identical potential differences at points, identical polarization potentials at the anode and cathode, and an identical potential drop across the electrolyte. The use of a best-estimate design eliminates multiple cycles of PS modeling. This approach enables the designer to exploit the advantages of both BE and PS modeling approaches.

Alloy Selection System for Elevated Temperatures The task of providing reliable information on corrosion of metals and alloys in high-temperature gases is difficult, due to the diversity of corrosive environments and the lack of standardization for corrosion data. The Alloy Selection System for Elevated Temperatures (ASSET) information system is being developed to address many of these difficulties. The present system contains information on approximately 80 commercial alloys,

4900 corrosion data measurements, and over six million exposure hours. Data compilation in this interactive model has been organized to allow prediction of sound metal thickness losses by several corrosion mechanisms at high temperatures as functions of gas composition, temperature, time, and alloy type (Ref 23, 24). The equations, which correlate the corrosion measurements with exposure conditions and the data, are stored in databases. The corrosion mechanisms for which corrosion predictions can be made are: • • • •

Sulfidation Sulfidation/oxidation Isothermal oxidation Carburization

The program stores the corrosion measurements, exposure conditions, and corrosion mechanisms and predicts alloy corrosion by accessing the stored data for that alloy and determining the parameters of the rate equation. A different equation is used for each corrosion mechanism. The software uses the alloy composition and the corrosive environment information to calculate the stable corrosion products and the equilibrium gas composition for a given combination of alloy and exposure conditions. The computations use the ChemSage program from FACT, a Gibbs free energy minimization program (Ref 25). Some of the potential solution or mixed corrosion product phases considered are MO, M3O4, M2O3, MS, M2S4, M3C, M23C6, and M7C3, where M represents a combination of iron, nickel, and chromium in the compound. The thermodynamic solution behaviors of the solid austenitic and ferritic alloys are also considered. Thermochemical characteristics such as the partial pressures and and the carbon activity of the environment, which help determine corrosion product stabilities, are also provided by the calculation. The software assists identification of the likely corrosion mechanism by knowing the stable corrosion products at the corrosion product/corrosive gas interface, the alloy in question, and the partial pressures, and . Different alloys in the same exposure conditions may exhibit different stable corrosion products and different corrosion mechanisms. In the absence of experimental data for the specific conditions of interest, predictions made by using this approach may be the best available for the corrosion mechanisms that are incorporated into the system, in comparison with those made using the traditional methods of literature review and data analysis. Examples of the accuracy using the ASSET system are shown in Fig. 9. They show how large amounts of corrosion data can be well correlated. The correlations are quite good for three decades of variation in corrosion penetration for several alloys and corrosion mechanisms, considering the uncertainty associated with these types of data.

Fig. 9 Verification of the corrosion prediction capabilities of ASSET over three decades of experimental corrosion rate measurements at various temperatures and in different environments with (a) carbon steel oxidation and (b) American Iron and Steel Institute (AISI) 310 sulfidation/oxidation. (Note: 1 mil equals 0.0254 mm.)

Predicting the Life of Nuclear Waste Containers In the United States and Canada, the regulations pertaining to the geologic disposal of highlevel nuclear waste require that the radionucleides remain substantially contained within the waste package for 300 to 1000 years after permanent closure of the repository. The current concept of a waste package involves the insertion of spent fuel bundles inside a container that is then placed in a deep borehole, either vertically or horizontally, with a small air gap between the container and the borehole. For vitrified wastes, a pour canister inside the outer container acts as an additional barrier, making the successful performance of the container material crucial to fulfilling the containment requirements over long periods of time. Providing that no failures occur as a result of mechanical effects, the main factor limiting survival of these containers is expected to be corrosion in the groundwater to which they would be exposed. Two general classes of container materials have been studied internationally: corrosion-allowance and corrosion-resistant materials. Corrosion-allowance materials possess a measurable general corrosion rate but are not susceptible to localized corrosion. By contrast, corrosion-resistant materials are expected to have very low general corrosion rates because of the presence of a protective surface oxide film. However, they may be susceptible to localized corrosion damage. A model developed to predict the failure of grade 2 titanium has considered two major corrosion modes (Ref 26): • •

Failure by crevice corrosion Failure by hydrogen-induced cracking (HIC)

In this predictive model, a small number of containers are assumed to be defective and to fail within 50 years of emplacement. Each modeling parameter is assigned a range of values, resulting in a distribution of corrosion rates and failure times. The crevice corrosion rate is assumed to be solely dependent on properties of the material and the temperature of the vault. Crevice corrosion is also assumed to initiate rapidly on all containers and subsequently propagate without repassivation. Failure by HIC is assumed to be inevitable once a container temperature falls below 30 °C (86 °F). However, the concentration of atomic hydrogen to render a container susceptible to HIC would occur only very slowly and might even be negligible if that container had never been subject to crevice corrosion. Figure 10 illustrates the thin-shell packed-particulate design chosen as a reference container for this study. The mathematical procedure to combine various probability functions and arrive at a failure probability of a hot container due to crevice corrosion at a certain temperature is illustrated in Fig. 11. The failure rate due to HIC in this model is arbitrarily assumed to have a triangular distribution in order to simplify the calculations and knowing that HIC is only a predicted marginal failure mode in the considered burial conditions. On the basis of these assumptions and the subsequent calculations, it was predicted that 96.7% of all containers would fail by crevice corrosion and the remaining by HIC. However, only 0.137% of the total number of containers was predicted to fail before 1000 years (0.1% by crevice corrosion and 0.037% by HIC), with the earliest failure after 300 years (Ref 26). For a view of nickel alloys in similar applications, see the article “Effects of Metallurgical Variables on the Corrosion of High-Nickel Alloys” in this Volume.

Fig. 10 Packed-particulate supported-shell container for waste nuclear fuel bundles

Fig. 11 Procedure used to determine the failure rate of hot containers as a function of time

Pitting Corrosion Fatigue (PCF) Models The phases of life of a structure have been classified as follows (Ref 27): • • • •

Phase 1: nucleation or formation of damage by a specific physical or corrosion damage process. Corrosion and other processes may act alone to create the damage. Phase 2: microstructurally dominated crack linkup and propagation Phase 3: crack propagation in the regime where various fracture mechanics methods may be applied, both for analysis and material characterization Phase 4: final instability of the system

Figure 12 illustrates these degradation processes, with the portion of life on the abscissa and the corresponding growth in discontinuity size plotted schematically on the ordinate. In a first model to estimate the number of fatigue cycles for a pit to reach the critical pit depth to nucleate a mode I crack, it was proposed that pit growth rate theory applied to data obtained with fatigue crack growth experiments in a corrosive environment would be sufficient. Using this model, the pit-to-crack transition length and cycles to failure for various stresses could be determined. However, there are many unknowns for the analysis of real components to estimate accurately the fatigue life under PCF conditions (Ref 27). Once critical-sized pits are formed, it is necessary to estimate the time or cycles for the pits to reach a critical condition or critical depth to nucleate fatigue cracks from those pits. Pit-to- crack transition length under various stresses using the stress-intensity threshold value from fatigue crack growth experiments can be estimated in a two-step process (Ref 27): 1. Pit-to-crack transition analysis 2. Estimation of fatigue cycles to failure

Fig. 12 Four phases involved in the degradation process of in-service components, where “A” corresponds to the first detectable crack, and the crack length progression steps correspond to (1) nucleation, (2) phase steps due to local anisotropy, (3) stress-dominated crack growth, and (4) crack to fracture. Source: Ref 27 The pit-to-crack transition analysis is first performed to determine the critical size of a pit in terms of pit depth that would transition to a crack (not necessarily a mode I crack) for different stresses. The stress values that could be used in the calculation are the estimated maximum applied stress that the component would be subjected to and the ultimate stress for the material in question. The PCF model can be used to determine the critical pit depth (Ref 27): (Eq 6) where, Ksf is the stress-intensity factor for a surface discontinuity (MPa/m0.5), σ is the applied stress (MPa), a is the size of the pit in terms of pit depth (μm or m), and Q is the dimensionless shape parameter. In this calculation, it can be assumed that Ksf is equal to “short” crack stress-intensity threshold (ΔKscth) for the material. It is recommended that the value of ΔKscth be used, because the pit- to-crack transition first would result in a non- mode I crack, that is, in the short crack region. It is important to note that there is no standard value for the ΔKscth in the short crack region for a particular material, because there is no standard test method to measure the fatigue crack growth rates in this regime. Therefore, this value can either be determined from conducting short fatigue crack growth experiments or determined from literature for a specific material. The shape parameter, Q, for a surface crack can be assumed, depending on the pit morphology. For different stress levels, ranging from the estimated applied stress for the component to the ultimate stress of the material, the critical pit depth, a, that would enable the transition of the pit to a short crack can be determined. The resultant value of a can be compared with the measured depth of the pit from the failure analysis of a similar component, if there is any. Moreover, the calculated value of a can be correlated to the experimentally generated pit growth rate curve to estimate the phase 1 life. A model has been proposed to estimate allowable stresses based on the

allowable stress-intensity threshold (Ref 28). This model is expressed in Eq 7, where the allowable stress, Δσall, at which the particular component can be operated is determined from the allowable stress-intensity threshold, ΔKall, which itself is determined from corrosion fatigue experiments. The maximum pit depth, hmax, in Eq 7 is measured from corrosion pit growth rate experiments for a given machine-material-environment system: (Eq 7) where, ΔKall can be determined from a fatique crack growth rate, da/dN, versus stress-intensity factor, ΔK, plot for a material; hmax is the maximum pit depth; and F is a geometric factor. The total cycles to failure under corrosion fatigue conditions can subsequently be estimated using the well-known Paris equation: da/dN = CΔKn

(Eq 8)

where, da/dN is the rate of crack growth per cycle (m/cycle), C and n are empirical parameters, and ΔK is the stress-intensity range (MPa · m0.5) The accuracy of the estimation can be improved if fatigue crack growth rate data under realistic corrosive environments are used in calculating the total cycles to failure. As well, the data on the effect of prior corrosion on the fatigue crack propagation also can help in getting more accurate estimation. The reader is referred to Ref 27 for a detailed example of the applicability of the PCF model in estimating the fatigue life of a real component.

References cited in this section 3. G.E.P. Box and N.R. Draper, Empirical Model-Building and Response Surfaces, John Wiley and Sons, 1987 4. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, NACE International, 1974 5. D.R.H. Jones, Materials Failure Analysis: Case Studies and Design Implications, Pergamon Press, 1993 6. P.R. Roberge, Handbook of Corrosion Engineering, McGraw Hill, 1999 7. W.T. Thompson, M.H. Kaye, C.W. Bale, and A.D. Pelton, Pourbaix Diagrams for Multielement Systems, Uhlig's Corrosion Handbook, R.W. Revie, Ed., Wiley Interscience, 2000, p 125–136 8. C. Wagner and W. Traud, Uber die deutung von korrosionvorgangen durch uberlagerung von elektrochemischen teilvorgangen und uber die potentialbildung an mischellektroden, Z. Elektrochemie, Vol 44, 1938, p 391-402 9. U.R. Evans, An Introduction to Metallic Corrosion, Edward Arnold, London, U.K., 1948 10. J.I. Munro, Anodic Protection of White and Green Liquor Tankage With and Without the Use of Protective Organic Linings, Proceedings of the Seventh International Symposium on Corrosion in the Pulp and Paper Industry, (Atlanta, GA), Technical Association of the Pulp and Paper Industry (TAPPI), 1992, p 117–130 11. D. Tromans, Anodic Polarization Behavior of Mild Steel in Hot Alkaline Sulfide Solutions, J. Electrochem. Soc., Vol 127, 1980, p 1253–1256 12. A. Anderko, P. McKenzie, and R.D. Young, Computation of Rates of General Corrosion Using Electrochemical and Thermodynamic Models, Corrosion, Vol 57, 2001, p 202–213 13. A. Anderko and R.D. Young, “A Model for Calculating Rates of General Corrosion of Carbon Steel and 13% Cr Stainless Steel in CO2/H2S Environments,” Paper 086, Corrosion 2001, NACE International, 2001

14. A. Anderko and R.D. Young, A Model for Corrosion of Carbon Steel in Lithium Bromide Absorption Refrigeration Systems, Corrosion, Vol 56, 2000, p 543–555 15. M. Rafal, J.W. Berthold, N.C. Scrivner, and S.L. Grise, Models for Electrolyte Solutions, Models for Thermodynamic and Phase Equilibria Calculations, S.I. Sandler, Ed., Marcel Dekker, 1995, p 601–670 16. A.H. Truesdell and B.F. Jones, A Computer Program for Calculating Chemical Equilibria of Natural Waters, J. Res., US Geol. Surv., Vol 2, 1974, p 233–248 17. R.J. Ferguson, A.J. Freedman, G. Fowler, A.J. Kulik, J. Robson, and D.J. Weintritt, The Practical Application of Ion Association Model Saturation Level Indices to Commercial Water Treatmant Problem Solving, Mineral Scale Formation and Inhibition, Z. Amjad Ed., Plenum Press, 1995, p 323– 340 18. R.J. Ferguson, O. Codina, W. Rule, and R. Baebel, “Real-Time Control of Scale Inhibitor Feed Rate,” IWC-88-57, 49th Annual Meeting, (Pittsburg, PA). International Water Conference, 1988 19. R.J. Ferguson, “Computerized Ion Association Model Profiles Complete Range of Cooling System Parameters,” IWC-91-47, 52nd Annual Meeting, (Pittsburg, PA), International Water Conference, 1991 20. R.A. Adey and J. Baynham, “Design and Optimization of Cathodic Protection Systems Using Computer Simulation,” Paper 723, Corrosion 2000, NACE International, 2000 21. V.G. DeGiorgi and K.E. Lucas, Computational Design of ICCP Systems: Lessons Learned and Future Directions, Designing Cathodic Protection Systems for Marine Structures and Vehicles, STP 1370, H.P. Hack, Ed., ASTM, 1999, p 87–100 22. R.W. Ditchfield, J.N. McGrath, and D.J. Tighe-Ford, Theoretical Validation of the Physical Scale Modeling of the Electrical Potential Characteristics of Marine Impressed Current Cathodic Protection, J. Appl. Electrochem., Vol 25, 1995, p 54–60 23. R.C. John, A.D. Pelton, A.L. Young, W.T. Thompson, I.G. Wright, and T.M. Besmann, Managing Corrosion Engineering Information for Metals in Hot Corrosive Gases, Materials Design Approaches and Experiences, J.C., Zhao, M. Fahrmann, and T.M. Pollock, Ed., 4–8 Nov 2001 (Indianapolis, IN), TMS, 2001, p 359–382 24. R.C. John, “Compilation and Use of Corrosion Data for Alloys in Various High- Temperature Gases,” Paper 99073, Corrosion 99, NACE International, 1999 25. C.W. Bale, A.D. Pelton, and W.T. Thompson, Facility for the Analysis of Chemical Thermodynamics (F*A*C*T), Ecole Polytechnique/McGill University, Montreal, Canada, 1996 26. D.W. Shoesmith, B.M. Ikeda, and D.M. LeNeveu, Modeling the Failure of Nuclear Waste Containers, Corrosion, Vol 53, 1997, p 820–829 27. D.W. Hoeppner, V. Chandrasekaran, and A.M.H. Taylor, “Review of Pitting Corrosion Fatigue Models,” ICAF '99, 12–16 July 1999 (Bellevue, WA), International Committee of Aeronautical Fatigue (ICAF), 1999 28. S. Kawai and K. Kasai, Considerations of Allowable Stress of Corrosion Fatigue (Focused on the Influence of Pitting), Fatigue Fract. Eng. Mater. Struct., Vol 8, 1985, p 115–127

P.R. Roberge, Modeling Corrosion Processes, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 430–445 Modeling Corrosion Processes P.R. Roberge, Royal Military College of Canada

Risk-Based Models Some of the issues involved in deciding on a cost-effective solution for combating corrosion are generic to sound system management. Others are specifically related to the impact of corrosion damage on system integrity and operating costs. Risk assessments can be designed to answer many questions. Some have risk hierarchies with a limited number of variables to produce a quick screening of high-risk system components. Others consider the impact of a multitude of variables in their overall assessment. The danger in using too few variables is that the analysis becomes more qualitative than quantitative, approaching more of the fundamental indexing method. Similarly, when considering a large number of variables, the potential exists for diluting the impact of rare-occurrence events among the myriad of other potential problems. Therefore, the development and fine-tuning of the risk algorithm to ensure that potentially dangerous conditions are properly identified is the most critical step in any risk analysis program (Ref 29).

Basic Principles of Risk Management Traditional risk management has been based on regulatory code compliance issues, and the decision to perform maintenance has been left up to a limited number of people, drawing off of their own experience and the limited amount of information they can access. Risk-based models use a systematic and structured approach based on the consequences of failure. As such, they represent a shift away from time-based tasks and emphasize the functional importance of system components and their failure/maintenance history. The following examples illustrate how these considerations are put into practice and integrated into efficient management systems. Corrosion-Based Design (CBD) Applied to a Steam Generator. More industrial equipment is being designed or forced to perform for durations for which the environmental effects have never been seriously considered. An objective of the CBD method is to provide a framework within which materials and design disciplines can communicate effectively. The basic principles for conducting CBD are (Ref 30): •

•

•

All engineering materials are simply reactive chemicals that, in some environments, dissolve as rapidly as table salt in hot water. The surprise is never that materials fail; the surprise should be that these reactive materials work. Materials fail in the environments that are specifically adjacent to the surface and not in the bulk environment. These local environments are greatly influenced by prior history of exposure, flow, buildup of deposits, and local cells. What appears to be imponderable complexity in corrosion can be handled and bounded by explicit consideration of specific possible environments and modes of corrosion that can occur.

The CBD approach to predicting performance is a series of knowledge-elicitation steps that require detailed consideration on a multitude of topics. The two most important of these steps are shown in Fig. 13 and 14, which describe, respectively, the environment and the material definitions. Each of the numbers in brackets in Fig. 13 identifies an explicit action that needs to be considered in the step of environmental definition. The endpoint of the process is an input to a location-for-analysis (LA) matrix, whose locations in a steam generator are indicated in Fig. 15. A brief explanation of the individual elements in Fig. 13 follows: 1. Nominal chemistry refers to the bulk chemistry. For components exposed to ordinary air atmospheres, the major elements mean humid air. The minor elements refer to industrial contaminants, such as SO 2 and NO2.

2. Prior chemistry history refers to exposures to environmental species that might still reside on the surfaces or inside crevices. 3. System sources refer to those environments that do not come directly from a component but from outside. 4. Physical features include occluded geometries, flow, and long-range electrochemical cells. 5. Transformations refer, for example, to microbial actions that can change relatively innocuous chemicals, such as sulfates, into very corrosive sulfide species that accelerate hydrogen entry and increase corrosion rates. 6. Concentration refers to accumulations much greater than that in the bulk environment due to various actions of wetting and drawing, evaporation, potential gradients, and crevice actions that prevent dilution. 7. Inhibition refers to actions taken to minimize corrosive actions. This usually involves additions of oxygen scavengers or species that interfere directly with anodic or cathodic reactions.

Fig. 13 Analysis sequence for determining environment at a location-for-analysis (LA) matrix. Details are given in text.

Fig. 14 Analysis sequence for determining materials at a location-for-analysis (LA) matrix. Details are given in text.

Fig. 15 Schematic view of steam generator indicating the locations for analysis considered in the template presented in Table 4 Defining materials characterizes those aspects that affect the modes of corrosion. A brief explanation of the individual elements in Fig. 14 follows: 1. Major nominal alloy composition includes those species added in relatively large concentrations, such as the iron, chromium, nickel, and molybdenum in type 316 stainless steel. 2. Minor nominal alloy composition includes species such as sulfur, nitrogen, phosphorus, silicon, and manganese usually found in steels, or the oxygen and nitrogen usually found in titanium. 3. Impurities include species such as arsenic, selenium, and similar elements that contribute to temper brittleness, or to excess oxygen and nitrogen picked up during welding of titanium. 4. Processing includes fabrication to shape, with associated times and temperatures of hot deformation and heat treatments, cold deformation to size, welding, machining, and grinding. 5. Structure includes single and multiple phases, anisotropy, second phases, and grain size. 6. Embrittlement includes grain-boundary composition and the formation of phases that embrittle, such as sigma phases in iron-chromium-base alloys. 7. Surface includes surface machining, surface stresses that result from thermal gradients, grinding, shot peening, and impurities absorbed at the surface. 8. Mechanical includes the mechanical properties of tensile strength, ductility, and toughness.

The LA template of the locations that correspond to the most likely failure sites along tubes in a steam generator of a pressurized-water nuclear power plant, indicated in Fig. 15, is illustrated in Table 4 for the main failure modes and submodes considered in such an analysis.

Table 4 Matrix for organizing mode-location cells Locations for analysis (LAi)

Tubular expansion (i = 1) Tubular expansion (i = 2) Top of tube sheet (i = 3) Top of tube sheet (i = 4) Sludge (i = 5) Free span (i = 6) Free span (i = 7) Tube support (hot leg) (i = 8) Tube support (cold leg) (i = 9) U-bend (i = 10) U-bend AVB (i = 11)

ID OD Modes (MDj) and submodes (SDj) to be considered Submodes of SCC Submodes of IGC LPSCC HPSCC AcSCC MRSCC AkSCC PbSCC HPIGC AcIGC AkIGC Wastage Pitting Fatigue Wear (j = 1)

(j = 2)

(j = 3)

(j = 4)

(j = 5)

(j = 6)

(j = 7)

(j = 8)

(j = 9)

(j = 10)

…

…

…

…

…

…

…

…

…

… x

…

…

…

…

…

…

…

…

x

…

…

…

…

…

…

…

… x

…

…

…

…

…

…

… x

…

…

…

…

…

x

…

…

…

…

… x

…

…

…

… x

…

…

… x

…

x

x

…

…

…

…

… x

= (j = 12)

…

(j 11) …

…

(j = 13) …

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

…

ID, inside diameter; OD, outside diameter; SCC, stress-corrosion cracking; IGC, intergranular corrosion; LP, low potential; HP, high potential; Ac, acidic; MR, midrange pH; AK, alkaline; Pb, lead; AVB, antivibration bars A Fault Tree for the Risk Assessment of Gas Pipeline. A fault tree is a diagrammatic representation of the relationship between component-level failures and a system-level undesired event. A fault tree depicts how component-level failures propagate through the system to cause a system-level failure. The component-level failures are called the terminal events, primary events, or basic events of the fault tree. The system-level undesired event is called the top event of the fault tree. Figure 16 presents the most commonly used logic symbols and gates for the construction of fault trees (Ref 31). Fault tree analysis (FTA) is a very efficient method for reviewing and analytically examining a process or piece of equipment in its details to emphasize the lower-level fault occurrences that directly or indirectly contribute to a major fault or undesired event. By performing a FTA from lower-level failure mechanisms, a global overview of the system is achieved. Once completed, the fault tree allows the risk engineer to fully evaluate the safety or reliability of a system by altering the various lower-level attributes of the tree. Through this type of modeling, a number of variables may be visualized in a cost-effective manner.

Fig. 16 Fault tree symbols for gates, transfers, and events Figure 17 illustrates how Nova Corporation adopted FTA for the risk assessment of uniform corrosion on its 18,000 km (11,000 mile) gas pipeline network (Ref 32). The Nova rupture- risk FTA was normally performed for the review and analytical examination of systems or equipment to emphasize the lower-level fault occurrences, and the results of the FTA calculations were regularly validated with inspection results. These results also served to schedule maintenance operations, conduct surveys, and plan research and development efforts. Each element of each branch in Fig. 17 contains numeric probability information related to technical and historical data for each segment of the complete pipeline network. In some cases, it is simpler to assume some probability values for an entire system. The probabilities of operating at maximum permitted pressure and the presence of electrolyte were both set at value unity in Fig. 17, therefore forcing the focus on worst-case scenarios. Other, more verifiable variables can be fully developed, as is shown in Fig. 18 for two basic events describing the probable impact of a cathodic protection deficiency on the pipeline network.

Fig. 17 Fault tree for natural gas pipeline outage due to general corrosion. C.P., cathodic protection

Fig. 18 Detailed code for the basic events leading to a cathodic protection (C.P.) deficiency Maintenance Steering Group (MSG) System Applied to Aircraft. The aircraft industry and its controlling agencies have developed, over the past few decades, a comprehensive system to represent potential failures of aircraft components, the maintenance steering group (MSG). The first generation of formal air carrier maintenance programs was based on the belief that each part on an aircraft required periodic overhaul. As experience was gained, it became apparent that some components did not require as much attention as others, and new methods of maintenance control were developed. Condition monitoring was thus introduced in the decision logic of the initial MSG document (MSG-1) and was applied to Boeing 747 aircraft. After several alterations to the system, industry and regulatory authorities began working together in 1991 to provide additional enhancements to MSG-3. As a result of these efforts, revision 2 was submitted to the Federal Aviation Administration (FAA) in September 1993 and was accepted a few weeks later (Ref 33). One particularly interesting addition in this revision is a procedure for incorporating corrosion damage in the MSG logic. The environmental deterioration analysis (EDA) involves the evaluation of the structure against probable exposure to adverse environments. The evaluation of deterioration is based on a series of steps supported by reference materials containing baseline data expressing the susceptibility of structural materials to various types of environmental damage, and it produces inputs for the structure maintenance program. The MSG-3 (revision 2) structures analysis begins by developing a complete breakdown of the aircraft systems to the component level. All structural items are then either classified as structure-significant items (SSIs) or other structure. An item is classified as a SSI depending on consideration of consequences of failure and likelihood of failures as well as material, protection, and probable exposure to corrosive environments. All SSIs are then listed and categorized as damage-tolerant or safe-life items, to which life limits are assigned (Ref 34). Accidental damage, environmental deterioration, corrosion prevention and control, and fatigue damage evaluation are then performed for these SSIs, following the logic diagram illustrated in Fig. 19. Once the MSG3 structure analysis is completed, each element of the structural analysis diagram (Fig. 19) can be expanded right to the individual components and associated inspection and maintenance tasks. The logic of the EDA,

illustrated in Fig. 20, requires the input of a multitude of parameters, such as component part number, location, material composition, and protective coating, that can be actuated in practical templates.

Fig. 19 Overall maintenance steering group (MSG-3), revision 2, structural analysis logic diagram for aircraft. SSIs, structure-significant items

Fig. 20 Environmental deterioration analysis logic diagram. NDI, nondestructive inspection Corrosion Index for Pipeline Risk Evaluation. The third example of comprehensive models is the pipeline risk assessment methodology, described in detail in (Ref 35). The methodology is based on subjective risk assessment, a method particularly well adapted to situations where knowledge is incomplete and judgments are based on opinions, experience, and other nonquantifiable resources. The technique used for quantifying risk factors is a hybrid of several methods, allowing the user to confide in scores obtained by combining statistical failure data with operator experience. The subjective scoring system allows the assessment of the pipeline risk picture in two general parts: •

•

A detailed itemization and relative weighting of all reasonably foreseeable events that may lead to the failure of a pipeline are carried out. The itemization is further broken down into the four indexes illustrated in Fig. 21, corresponding to typical categories of pipeline accident failures. By considering each item in each index, an expert evaluator arrives at a numerical value for that index. The four index values are then summed to obtain the total index value. A detailed analysis is made of potential consequences of each failure, considering product characteristics, pipeline operating conditions, and line location.

Fig. 21 Pipeline risk assessment model The risk of a pipeline failure caused by corrosion, directly or indirectly, is probably the most common hazard associated with steel pipelines. The corrosion index was organized into three categories to reflect three types of environment to which pipelines are typically exposed, that is, atmospheric corrosion, soil corrosion, and internal corrosion. Table 5 contains the elements contributing to each type of environment and the suggested weighting factors. Building the risk assessment tool requires the following four steps: 1. Sectioning: dividing a system into smaller sections. The size of each section should reflect practical considerations of operation, maintenance, and cost of data gathering versus the benefit of increased accuracy. 2. Customizing: deciding on a list of risk contributors and risk reducers and their relative importance 3. Data gathering: building the database by completing an expert evaluation for each section of the system 4. Maintenance: identifying when and how risk factors can change and updating these factors accordingly Table 5 Pipeline corrosion risk subjective assessment Corrosion problem Atmospheric corrosion Facilities Atmospheric type Coating/inspection

Risk weight, points 0–5 0–10 0–5

0–20 Total Internal corrosion 0–10 Product corrosivity 0–10 Internal protection 0–20 Total Soil corrosion 0–8 Cathodic protection 0–10 Coating condition 0–4 Soil corrosivity 0–3 Age of system 0–4 Other metals AC-induced currents 0–4 0–5 SCC and HIC 0–6 Test leads Close internal surveys 0–8 0–8 Inspection tool 0–60 Total Total 0–100 AC, alternating current; SCC, stress-corrosion cracking; HIC, hydrogen- induced cracking

References cited in this section 29. S.F. Biagiotti, Jr. and M.P. Gloven, “Pipeline Risk Assessment and Effective Resource Allocation,” Paper 481, Corrosion 2000, NACE International, 2000 30. R.W. Staehle, Lifetime Prediction of Materials in Environments, Uhlig's Corrosion Handbook, R.W. Revie, Ed., Wiley Interscience, 2000, p 27–84 31. Fault Tree Analysis Application Guide, Reliability Analysis Center, 1990 32. P.R. Roberge, Eliciting Corrosion Knowledge Through the Fault-Tree Eyeglass, Modelling Aqueous Corrosion: From Individual Pits to Corrosion Management, K.R. Trethewey and P.R. Roberge, Ed., Kluwer Academic Publishers, 1994, p 399–416 33. D. Nakata, MSG-3 Aircraft Systems/Powerplant Analysis Method, Aircraft Maintenance and Reliability Seminar/Workshop, Vol 1, Transportation Systems Consulting Corp., 1997 34. BBAD Maintenance Program: Policy and Procedures Handbook, Bombardier Inc., 1994 35. W.K. Muhlbauer, Pipeline Risk Management Manual, Gulf Publishing Co., 1996


Knowledge Models

The adequate transfer and reuse of information involves the development of information- processing strategies that can become quite complex. Humans are accustomed to working with imprecise information. They naturally accept vague use of language, making continuous interpretations of the information they receive based on context. Ambiguity can cause problems. However, if the scope for ambiguity is provided, a question and answer clarifies the matter. For a machine to carry out the same tasks, unambiguous language must be defined. This, of course, is the first event in the creation of a traditional programming language, such as FORTRAN, PASCAL or C. The structuring of knowledge inevitably requires the definition of terms, statements of relationships between them, and elimination of ambiguity. Despite this, modern mathematics has devised techniques that do allow for imprecise or error-containing information, and, by means of fuzzy logic, some computers can now cope with such situations (Ref 36). Some of the artificial intelligence tools that have been developed in support of corrosion control and protection are reviewed in the following sections.

Basic Principles of Knowledge Models From a practical point of view, all models of expertise have their place and possess special features designed to represent specific types of knowledge. Knowledge engineering research has produced a wide range of models of expertise, which have been classified into five major classes (Ref 37): •

•

•

•

•

Heuristic models: offer a symbolic, or representational, description of expertise that is seen as a combination of facts and heuristics. A rule-based or heuristic system would be an excellent choice, for example, when the hierarchy between facts and decisions is well established. Deep models: suggest that experts use a variety of deep knowledge structures represented differently from heuristic rules. Deep models of expertise are said to better represent the expert functions at each level of the expertise model hierarchy. Implicit models: offer a representational description with a particular emphasis on the implicitness or explicitness of knowledge. Implicit models have become increasingly popular, with the development of artificial neural networks. Competence models: are representation independent and principally distinguish between domain and task knowledge. The focus of competence models is usually on the strategy of a problem-solving process. These models provide an extension of a functional rule-based system from which the heuristic knowledge was extracted to generate an explicit knowledge-level description of a specialist's expertise (Ref 37). Distributed models: can be compared to negotiation schemes. The tools implementing this approach are still on the research bench, because many aspects of fuzzy systems development for fuzzification and defuzzification remain both attractive and mysterious to apply (Ref 36).

Expert Systems. During the 1970s, research in expert systems (ES) was mostly a laboratory curiosity. The research focus then was really centered on developing ways of representing and reasoning about knowledge in a computer and much less on designing actual systems (Ref 38). In 1985, approximately 50 systems were deployed and reported, but the success of some of these captured the attention of many organizations and individuals. The corrosion community reacted with interest to the advent of these new information-processing technologies by establishing programs to foster and encourage the introduction of ES in the workplace. While some of these programs were relatively modest, others were quite ambitious and important both in scope and in funding. The advantages and limitations of using ES technology were analyzed in great detail in one of the first reported efforts on combating corrosion with ES (Ref 39). The stress-corrosion cracking (SCC) ES (SCCES) had been created to calculate the risk of various factors involved in SCC, such as crack initiation, when the user supplied evidence. The main goal of this effort was to support the decision process of “general” materials engineers. The system would initially play the role of a consultant, but it was anticipated that SCCES had the potential to become: • • • •

An intelligent checklist A trainer An expert sharpener A communication medium

•

A demonstration vehicle

The transfer of corrosion expertise into ES has been realized in a multitude of projects. The NACE International Conference Proceedings, for example, regularly contain papers that illustrate the continuous interest in the application of knowledge engineering to corrosion. Unfortunately, many systems reported in the literature have never been commercialized. This has resulted in a lack of impartial and practical information concerning the performance and accuracy of these systems. It is indeed very difficult to believe everything that is said in a paper or, even more so, when the information is some form of publicity. To remedy this situation, the European Federation of Corrosion (EFC) and the Materials Technology Institute (MTI) have performed two surveys, between 1988 and l990, requesting recognized developers of ES in corrosion-related areas to provide very specific information concerning the availability, scope, and performance of their systems (Ref 40). A review of the information presented in these reports was summarized (Ref 41), and other developments in expert systems for corrosion prevention and control were reported subsequently (Ref 42). Neural Networks. A modeling neural network essentially acts as a mapping operator or a transfer function, taking inputs normally fed to a process and computing its predicted output values (Ref 36). Because the inputoutput mapping can be either static or dynamic, the technology can be applied to a broad range of applications, being particularly efficient with real-time operations. However, neural networks are also well adapted to perform other implicit expert functions, such as pattern recognition and classification. An artificial neural network (ANN) is a network of many very simple processors or neurons (Fig. 22), each possibility having a small amount of local memory. The interaction of the neurons in the network is roughly based on the principles of neural science. Unidirectional channels that carry numeric data based on the weights of connections connect these neurons, which operate only on their local data and on the inputs they receive via the connections. Most neural networks have some sort of training rule. The training algorithm adjusts the weights on the basis of presented patterns. In other words, neural networks “learn” from examples. Artificial neural networks excel particularly at problems where pattern recognition is important and precise computational answers are not required. When ANNs inputs and/or outputs contain evolved parameters, their computational precision and extrapolation ability significantly increases and can even outperform more traditional modeling techniques. Some applications of ANN to corrosion problems are briefly described below.

Fig. 22 Schematic of a single processor or neuron in an artificial neural network Predicting the SCC Risk of Stainless Steels. The risk of encountering a SCC situation was functionalized in terms of the main environment variables (Ref 43). Case histories reflecting the influence of temperature, chloride concentration, and oxygen concentration have been analyzed by means of a back-propagation network. Three neural networks were developed. One was created to reveal the temperature and chloride concentration dependency (Fig. 23) and another to expose the combined effect of oxygen and chloride content in the environment. The third ANN was trained to explore the combined effect of all three parameters. During this project, ANNs were found to outperform traditional mathematical regression techniques, where the functions have to be specified before performing the analysis.

Fig. 23 Four-layer neural network architecture (one input, one output, and two middle computing hidden layers) for the prediction of stress-corrosion cracking (SCC) risk of austenitic stainless steels in industrial processes Corrosion Prediction from Polarization Scans. An ANN was put to the task of recognizing certain relationships in potentiodynamic polarization scans in order to predict the occurrence of general or localized corrosion, such as pitting and crevice (Ref 44). The initial data inputs were derived by carefully examining a number of polarization scans for a number of systems and recording those features that were used for the predictions. Table 6 lists the initial inputs used and how the features were digitized for computer input. The variables shown were chosen because they were thought to be the most significant in relation to the predictions (Table 6). The final ANN proved to be able to make appropriate predictions using scans outside the initial training set. The resulting ANN was embedded in an ES to facilitate the input of data and the interpretation of the numerical output of the ANN. Table 6 Data inputs and outputs for predicting corrosion out of polarization scans with an artificial neural network Parameter Input Repassivation potential Pitting potential Hysteresis

Current density at scan reversal Anodic nose Passive current density Potential at anodic-cathodic transition Output Crevice corrosion predicted Pitting predicted

Value of feature Eprot - Ecorr Epit - Ecorr +1 = Positive 0 = None -1 = Negative μA/cm2 +1 = Yes 0 = No μA/cm2 Ea to c - Ecorr +1 = Yes 0 = No +1 = Yes

0 = No Should general corrosion be considered +1 = Yes 0 = No Eprot, repassivation (or protection) potential; Ecorr, corrosion potential; Epit, pitting potential; EA to C, anodic to cathodic potential Predicting the Degradation of Nonmetallic Lining Materials. An ANN was trained to recognize the pattern between results from a sequential immersion test for nonmetallic materials and the behavior of the same material in field applications (Ref 45). Eighty-nine cases were used for the supervised training of the network. Another 17 cases were held back for testing of the trained network. An effort was made to ensure that both sets had experimental data taken from the same test but using different samples. Appropriate choice of features enabled the ANN to mimic the expert with reasonable accuracy. The successful development of this ANN was another indication that ANNs could seriously aid in projecting laboratory results into field predictions. Validation and Extrapolation of Electrochemical Impedance Data. The ANN developed in this project had three independent input vectors: frequency, pH, and the applied potential (Ref 46). The ANN was designed to learn from the invisible or hidden information at high and low frequencies and to predict in a lower frequency range than that used for training. Eight sets of impedance data acquired on nickel electrodes in phosphate solutions were used for this project. Five sets were used for training the ANN and three for its testing. The ANN proved to be a powerful technique for generating diagnostics at these conditions. Modeling CO2 Corrosion. A CO2 corrosion worst-case model based on an ANN approach was developed and the model validated against a large experimental database (Ref 47). An experimental database was used to train and test the ANN. It consisted initially of six elemental descriptors (temperature, partial CO2 pressure, ferrous and bicarbonate ion concentrations, pH, and flow velocity) and one output, that is, the corrosion rate. The system demonstrated superior interpolation performance compared to two well-known semiempirical models. The ANN model also demonstrated extrapolation capabilities comparable to a purely mechanistic electrochemical CO2 corrosion model. Pitting of Aluminum in Natural Waters. The development and testing of an ANN model for the prediction of pitting of aluminum alloy A91100 was demonstrated in the context of nuclear waste storage (Ref 48). The U.S. Department of Energy Savannah River site for the extended storage of aluminum-clad spent nuclear fuel, where most of this waste is stored in large water basins, is fully equipped with corrosion control programs. Vigilant monitoring and regulation of water chemistry has reduced the occurrence of pitting to essentially zero. Trained ANN models, based on the configuration described in Fig. 24, were used to predict pit depths for periods up to 1 year. Because the supply waters contained species not included in the training data of these models, correlation between predicted and actual pit depth deteriorated for longer time periods.

Fig. 24 Four-layer neural network architecture (one input, one output, and two middle computing hidden layers) for the prediction of pitting of aluminum as a function of water chemistry variables

References cited in this section 36. K.R. Trethewey and P.R. Roberge, Design and Structure of Knowledge-Based Systems for Improved Materials Performance, Corros. Rev., Vol 15, 1997, p 71–94 37. P.E. Slatter, Models of Expertise in Knowledge Engineering, Knowledge Engineering, Vol 1, McGraw Hill, 1990 38. J. Durkin, Expert Systems: Design and Development, MacMillan Publishing Co., 1994 39. A. Basden, On the Applications of Expert Systems, Int. J. Man Mach. Stud., Vol 19, 1983, p 461–472 40. Report of Task Group 1 of the Working Party on Expert Systems in Materials Engineering, Materials Technology Institute, 1990, p 1–124 41. P.R. Roberge, Bridging the Gap Between the World of Knowledge and the World That Knows, Mater. Perform., Vol 33, 1994, p 52–56 42. P.R. Roberge, Expert Systems for Corrosion Prevention and Control, Corros. Rev., Vol 15, 1997, p 1– 14 43. H.M.G. Smets and W.F.L. Bogaerts, SCC Analysis of Austenitic Stainless Steels in Chloride-Bearing Water by Neutral Network Techniques, Corrosion, Vol 48, 1992, p 618–623 44. E.M. Rosen and D.C. Silverman, Corrosion Prediction from Polarization Scans Using Artificial Neural Network Integrated with an Expert System, Corrosion, Vol 48, 1992, p 734–745 45. D.C. Silverman, Artificial Neural Network Predictions of Degradation of Nonmetallic Lining Materials from Laboratory Tests, Corrosion, Vol 50, 1994, p 411–418 46. M. Urquidi-MacDonald and P.C. Egan, Validation and Extrapolation of Electrochemical Impedance Spectroscopy Data, Corros. Rev., Vol 15, 1997, p 169–194 47. S. Nesic and M. Vrhovac, A Neural Network Model for CO2 Corrosion of Carbon Steel, J. Corros. Sci. Eng., Vol 1, 1998, p 1–11 48. J. Leifer and J.I. Mickalonis, Prediction of Aluminum Pitting in Natural Waters via Artificial Neural Network Analysis, Corrosion, Vol 56, 2000, p 563–571


References

1. T. Kuhn, The Structure of Scientific Revolutions, University of Chicago Press, 1970 2. P. O'Rorke, S. Morris, and D. Schulenburg, Theory Formation by Abduction: A Case Study Based on the Chemical Revolution, Computational Models of Scientific Discovery and Theory Formation, J. Shrager and P. Langley, Ed., Morgan Kaufmann Publishing, 1990 3. G.E.P. Box and N.R. Draper, Empirical Model-Building and Response Surfaces, John Wiley and Sons, 1987 4. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, NACE International, 1974 5. D.R.H. Jones, Materials Failure Analysis: Case Studies and Design Implications, Pergamon Press, 1993 6. P.R. Roberge, Handbook of Corrosion Engineering, McGraw Hill, 1999 7. W.T. Thompson, M.H. Kaye, C.W. Bale, and A.D. Pelton, Pourbaix Diagrams for Multielement Systems, Uhlig's Corrosion Handbook, R.W. Revie, Ed., Wiley Interscience, 2000, p 125–136 8. C. Wagner and W. Traud, Uber die deutung von korrosionvorgangen durch uberlagerung von elektrochemischen teilvorgangen und uber die potentialbildung an mischellektroden, Z. Elektrochemie, Vol 44, 1938, p 391-402 9. U.R. Evans, An Introduction to Metallic Corrosion, Edward Arnold, London, U.K., 1948 10. J.I. Munro, Anodic Protection of White and Green Liquor Tankage With and Without the Use of Protective Organic Linings, Proceedings of the Seventh International Symposium on Corrosion in the Pulp and Paper Industry, (Atlanta, GA), Technical Association of the Pulp and Paper Industry (TAPPI), 1992, p 117–130 11. D. Tromans, Anodic Polarization Behavior of Mild Steel in Hot Alkaline Sulfide Solutions, J. Electrochem. Soc., Vol 127, 1980, p 1253–1256 12. A. Anderko, P. McKenzie, and R.D. Young, Computation of Rates of General Corrosion Using Electrochemical and Thermodynamic Models, Corrosion, Vol 57, 2001, p 202–213 13. A. Anderko and R.D. Young, “A Model for Calculating Rates of General Corrosion of Carbon Steel and 13% Cr Stainless Steel in CO2/H2S Environments,” Paper 086, Corrosion 2001, NACE International, 2001 14. A. Anderko and R.D. Young, A Model for Corrosion of Carbon Steel in Lithium Bromide Absorption Refrigeration Systems, Corrosion, Vol 56, 2000, p 543–555 15. M. Rafal, J.W. Berthold, N.C. Scrivner, and S.L. Grise, Models for Electrolyte Solutions, Models for Thermodynamic and Phase Equilibria Calculations, S.I. Sandler, Ed., Marcel Dekker, 1995, p 601–670 16. A.H. Truesdell and B.F. Jones, A Computer Program for Calculating Chemical Equilibria of Natural Waters, J. Res., US Geol. Surv., Vol 2, 1974, p 233–248 17. R.J. Ferguson, A.J. Freedman, G. Fowler, A.J. Kulik, J. Robson, and D.J. Weintritt, The Practical Application of Ion Association Model Saturation Level Indices to Commercial Water Treatmant Problem Solving, Mineral Scale Formation and Inhibition, Z. Amjad Ed., Plenum Press, 1995, p 323– 340

18. R.J. Ferguson, O. Codina, W. Rule, and R. Baebel, “Real-Time Control of Scale Inhibitor Feed Rate,” IWC-88-57, 49th Annual Meeting, (Pittsburg, PA). International Water Conference, 1988 19. R.J. Ferguson, “Computerized Ion Association Model Profiles Complete Range of Cooling System Parameters,” IWC-91-47, 52nd Annual Meeting, (Pittsburg, PA), International Water Conference, 1991 20. R.A. Adey and J. Baynham, “Design and Optimization of Cathodic Protection Systems Using Computer Simulation,” Paper 723, Corrosion 2000, NACE International, 2000 21. V.G. DeGiorgi and K.E. Lucas, Computational Design of ICCP Systems: Lessons Learned and Future Directions, Designing Cathodic Protection Systems for Marine Structures and Vehicles, STP 1370, H.P. Hack, Ed., ASTM, 1999, p 87–100 22. R.W. Ditchfield, J.N. McGrath, and D.J. Tighe-Ford, Theoretical Validation of the Physical Scale Modeling of the Electrical Potential Characteristics of Marine Impressed Current Cathodic Protection, J. Appl. Electrochem., Vol 25, 1995, p 54–60 23. R.C. John, A.D. Pelton, A.L. Young, W.T. Thompson, I.G. Wright, and T.M. Besmann, Managing Corrosion Engineering Information for Metals in Hot Corrosive Gases, Materials Design Approaches and Experiences, J.C., Zhao, M. Fahrmann, and T.M. Pollock, Ed., 4–8 Nov 2001 (Indianapolis, IN), TMS, 2001, p 359–382 24. R.C. John, “Compilation and Use of Corrosion Data for Alloys in Various High- Temperature Gases,” Paper 99073, Corrosion 99, NACE International, 1999 25. C.W. Bale, A.D. Pelton, and W.T. Thompson, Facility for the Analysis of Chemical Thermodynamics (F*A*C*T), Ecole Polytechnique/McGill University, Montreal, Canada, 1996 26. D.W. Shoesmith, B.M. Ikeda, and D.M. LeNeveu, Modeling the Failure of Nuclear Waste Containers, Corrosion, Vol 53, 1997, p 820–829 27. D.W. Hoeppner, V. Chandrasekaran, and A.M.H. Taylor, “Review of Pitting Corrosion Fatigue Models,” ICAF '99, 12–16 July 1999 (Bellevue, WA), International Committee of Aeronautical Fatigue (ICAF), 1999 28. S. Kawai and K. Kasai, Considerations of Allowable Stress of Corrosion Fatigue (Focused on the Influence of Pitting), Fatigue Fract. Eng. Mater. Struct., Vol 8, 1985, p 115–127 29. S.F. Biagiotti, Jr. and M.P. Gloven, “Pipeline Risk Assessment and Effective Resource Allocation,” Paper 481, Corrosion 2000, NACE International, 2000 30. R.W. Staehle, Lifetime Prediction of Materials in Environments, Uhlig's Corrosion Handbook, R.W. Revie, Ed., Wiley Interscience, 2000, p 27–84 31. Fault Tree Analysis Application Guide, Reliability Analysis Center, 1990 32. P.R. Roberge, Eliciting Corrosion Knowledge Through the Fault-Tree Eyeglass, Modelling Aqueous Corrosion: From Individual Pits to Corrosion Management, K.R. Trethewey and P.R. Roberge, Ed., Kluwer Academic Publishers, 1994, p 399–416 33. D. Nakata, MSG-3 Aircraft Systems/Powerplant Analysis Method, Aircraft Maintenance and Reliability Seminar/Workshop, Vol 1, Transportation Systems Consulting Corp., 1997 34. BBAD Maintenance Program: Policy and Procedures Handbook, Bombardier Inc., 1994

35. W.K. Muhlbauer, Pipeline Risk Management Manual, Gulf Publishing Co., 1996 36. K.R. Trethewey and P.R. Roberge, Design and Structure of Knowledge-Based Systems for Improved Materials Performance, Corros. Rev., Vol 15, 1997, p 71–94 37. P.E. Slatter, Models of Expertise in Knowledge Engineering, Knowledge Engineering, Vol 1, McGraw Hill, 1990 38. J. Durkin, Expert Systems: Design and Development, MacMillan Publishing Co., 1994 39. A. Basden, On the Applications of Expert Systems, Int. J. Man Mach. Stud., Vol 19, 1983, p 461–472 40. Report of Task Group 1 of the Working Party on Expert Systems in Materials Engineering, Materials Technology Institute, 1990, p 1–124 41. P.R. Roberge, Bridging the Gap Between the World of Knowledge and the World That Knows, Mater. Perform., Vol 33, 1994, p 52–56 42. P.R. Roberge, Expert Systems for Corrosion Prevention and Control, Corros. Rev., Vol 15, 1997, p 1– 14 43. H.M.G. Smets and W.F.L. Bogaerts, SCC Analysis of Austenitic Stainless Steels in Chloride-Bearing Water by Neutral Network Techniques, Corrosion, Vol 48, 1992, p 618–623 44. E.M. Rosen and D.C. Silverman, Corrosion Prediction from Polarization Scans Using Artificial Neural Network Integrated with an Expert System, Corrosion, Vol 48, 1992, p 734–745 45. D.C. Silverman, Artificial Neural Network Predictions of Degradation of Nonmetallic Lining Materials from Laboratory Tests, Corrosion, Vol 50, 1994, p 411–418 46. M. Urquidi-MacDonald and P.C. Egan, Validation and Extrapolation of Electrochemical Impedance Spectroscopy Data, Corros. Rev., Vol 15, 1997, p 169–194 47. S. Nesic and M. Vrhovac, A Neural Network Model for CO2 Corrosion of Carbon Steel, J. Corros. Sci. Eng., Vol 1, 1998, p 1–11 48. J. Leifer and J.I. Mickalonis, Prediction of Aluminum Pitting in Natural Waters via Artificial Neural Network Analysis, Corrosion, Vol 56, 2000, p 563–571

F. Mansfeld, Electrochemical Methods of Corrosion Testing, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 446–462

Electrochemical Methods of Corrosion Testing Florian Mansfeld, University of Southern California

Introduction DURING CORROSION, at least two electrochemical reactions, an oxidation and a reduction reaction, occur at a metal-electrolyte interface. Because corrosion is due to an electrochemical mechanism, it is clear that electrochemical techniques can be used to study corrosion reactions and mechanisms. This article reviews the fundamentals of electrochemical corrosion test methods, preparations for electrochemical polarization experiments, and the various electrochemical techniques available for laboratory studies of corrosion phenomena.

F. Mansfeld, Electrochemical Methods of Corrosion Testing, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 446–462 Electrochemical Methods of Corrosion Testing Florian Mansfeld, University of Southern California

Fundamentals The basis of all electrochemical techniques lies in the principle of mixed-potential theory, which was clearly established in a classic paper published in 1938 (Ref 1). It was demonstrated that uniform corrosion occurs when anodic and cathodic reactions take place in constant change of location and time of the individual processes. In this concept, it is not necessary to postulate the presence of local anodes and cathodes for corrosion to occur. Nevertheless, in later years it has often been stated that corrosion occurs when local anodes and cathodes have been established, despite the obvious fact that, in this case, uniform corrosion is not possible. The corrosion potential, Ecorr, or open-circuit potential (OCP), is a mixed potential; that is, it is the unique potential at which the rates of the anodic and cathodic reaction are exactly equal. The rate of the anodic or the cathodic reaction at Ecorr equals the corrosion rate. Because the rates of these two reactions are independent of each other, it is possible to inhibit corrosion by reducing the rate of only one of the two reactions by using, for example, anodic or cathodic corrosion inhibitors. Cathodic protection in which the material to be protected is polarized cathodically in order to reduce the rate of the anodic metal dissolution reaction is another example. The concept of mixed-potential theory and its application in corrosion science and technology has been discussed in many textbooks (Ref 2, 3). The use of mixed-potential theory for the monitoring of corrosion rates was also demonstrated by showing that the slope of a potential versus current density plot at Ecorr can be used to determine the corrosion rate (Ref 1). Today, this technique is referred to as the polarization- resistance technique or, less precisely, as the linear polarization technique (Ref 4). From this research, it also becomes clear that corrosion current densities, icorr, can be obtained from the analysis of anodic and cathodic polarization curves, which is often referred to as Tafel extrapolation (Ref 2, 3). This concept has been developed further (Ref 4, 5). In addition, mechanistic information can be obtained by analysis of the Tafel slopes that have characteristic values, depending on the nature of the rate-determining step in the corrosion mechanism (Ref 2, 4).

A large number of electrochemical techniques are available for laboratory studies of various corrosion phenomena as well as for corrosion monitoring (Ref 6, 7, 8, 9, 10). The National Association of Corrosion Engineers (NACE) series, Corrosion Testing Made Easy, contains several volumes that deal with the use of electrochemical methods (Ref 11, 12, 13). The advice given previously (Ref 14)—“Don't be afraid of electrochemical techniques, but use them with care”— applies for laboratory studies. Often, electrochemical techniques are considered to be overly complicated. However, recent developments by the manufacturers of hardware and software have made the use of these techniques quite simple. Most equipment contains software that enables the user to carry out an electrochemical test in a straightforward manner. Of course, interpretation of the results of such tests remains up to the user. Often, only the experimental results are presented; sometimes, the results of very simple measurements, such as Ecorr, have been overinterpreted. It has to be remembered that electrochemical tests indicate that corrosion has happened but do not contain information concerning the cause of corrosion. Therefore, statements that certain electrochemical techniques such as the electrochemical noise (ECN) technique “offers unique capabilities that should allow quantification of microbial effects on corrosion,” and “ECN offers promise in identifying microbiologically influenced corrosion” (Ref 15) should be viewed with caution. Electrochemical techniques available for laboratory studies of corrosion phenomena are summarized in Table 1. These techniques range from those in which no external signal is applied, such as Ecorr and electrochemical noise analysis (ENA), to those in which only a small potential or current perturbation is applied, such as polarization resistance, Rp, and electrochemical impedance spectroscopy (EIS); to those in which the applied potential or current is varied over a wide range, such as anodic and cathodic polarization curves and pitting scans. Table 1 Electrochemical corrosion testing techniques Category No applied signal

Test method Open circuit or corrosion potential Dissimilar metal corrosion (galvanic corrosion) Electrochemical noise analysis Polarization resistance (linear polarization) Small-signal polarization Electrochemical impedance spectroscopy Potentiostatic and galvanostatic polarization Large-signal polarization Potentiodynamic and galvanodynamic polarization Scanning electrode techniques Potential scans Current scans Electrochemical impedance spectroscopy scans Hydrogen permeation Miscellaneous tests Anodized aluminum corrosion test Electrolytic corrosion test Paint adhesion on a scribed surface Impedance test for anodized aluminumodized Critical pitting temperature The galvanic current flowing between two dissimilar metals can be monitored using a zero- resistance ammeter (ZRA). Diffusion of atomic hydrogen into metals, such as steel, that can cause failure can be evaluated with the electrochemical permeation method. Scanning electrode techniques can be used to determine the location of anodes and cathodes on electrodes. It has to be recognized that very simple techniques, such as those in which no external signal is applied, usually provide much less quantitative information concerning corrosion mechanisms than those in which a carefully controlled signal is applied to the system under investigation and the response of the system is analyzed. An example for this kind of approach is EIS, in which the frequency of the applied signal is used as an additional variable. ASTM Committee G-1 has published standard practices for most of the techniques listed in Table 1, describing the experimental approach and giving examples for the use of these methods.


1. C. Wagner and W. Traud, Z. Elektrochemie, Vol 44, 1938, p 391 2. H. Kaesche, The Corrosion of Metals, NACE, 1985 3. D.A. Jones, Principles and Prevention of Corrosion, MacMillan Publishing Co., 1992 4. F. Mansfeld, The Polarization Resistance Technique for Measuring Corrosion Currents, Advancements in Corrosion Science and Technology, Vol 6, M.G. Fontana and R.W. Staehle, Ed., Plenum Press, 1976, p 163 5. M. Stern and A.L. Geary, J. Electrochem. Soc., Vol 104, 1957, p 56 6. R. Baboian, Ed., Electrochemical Techniques for Corrosion, NACE, 1977 7. Electrochemical Corrosion Testing, STP 727, U. Bertocci and F. Mansfeld, Ed., ASTM, 1979 8. R. Baboian, Ed., Electrochemical Techniques for Corrosion Engineers, NACE, 1986 9. R. Baboian, Ed., Corrosion Tests and Standards—Application and Interpretation, ASTM, 1995 10. R.W. Revie, Ed., Uhlig's Corrosion Handbook, 2nd ed., Electrochemical Society Series, WileyInterscience, 2000 11. B.J. Little, P.A. Wagner, and F. Mansfeld, Corrosion Testing Made Easy—Microbiologically Influenced Corrosion, B.C. Syrett, Ed., NACE, 1997 12. N.G. Thompson and J.H. Payer, Corrosion Testing Made Easy—dc Electrochemical Methods, B.C. Syrett, Ed., NACE, 1998 13. R. Cottis and S. Turgoose, Corrosion Testing Made Easy—Electrochemical Impedance and Noise, B.C. Syrett, Ed., NACE, 1999 14. F. Mansfeld, Corrosion, Vol 44, 1988, p 856 15. G.G. Geesey, I. Beech, P.J. Bremer, B.J. Webster, and D.B. Wells, Biocorrosion, Biofilms II: Process Analysis and Applications, J.D. Bryers, Ed., Wiley, 2000, p 281


Preparing an Electrochemical Polarization Experiment For the performance of most electrochemical tests discussed here, a potentiostat and an electrochemical cell are required. Open-circuit-potential measurements may be performed using only a high-input impedance voltmeter, while galvanic current and ECN measurements may be carried out using a ZRA instead of a potentiostat, as

discussed subsequently. The features and requirements of the instrumentation needed for an electrochemical test are discussed in brief. A more detailed description can be found in Ref 12. The basic instrumentation needed for the electrochemical tests includes: • • • • • • • • • •

Test or working electrode (WE) One or more counter electrodes (CE) Reference electrode (RE) Test cell Potentiostat Voltmeters for monitoring of potential and current (optional) Recording devices (strip chart or x-y recorders) Computer with software program and plotter (replacing the voltmeter and recording devices) (optional) Gas tank with nitrogen, argon, or another gas to deaerate the solution (optional) Thermostat with a constant temperature bath or heating mantle (optional)

In a potentiostatic experiment, the applied potential, Eappl, is maintained between the RE and WE. The WE is at virtual ground. Therefore, the sign of the potential displayed on the potentiostat is sometimes that of the RE to ground and opposite the potential of the WE. This fact, unfortunately, often leads to confusion concerning the sign of the measured potential. The instruction manual of the potentiostat to be used should be consulted regarding this point. The current flows between CE and WE and is usually displayed with the correct sign (positive for an anodic current). A simple test can be carried out using a zinc WE, which has a very negative Ecorr versus any common RE, and for which a negative (cathodic) current is recorded when a potential E < Ecorr is applied. The minimum requirements for a potentiostatic test are a test cell with WE, RE, CE, a potentiostat, and a recording device (Ref 11, 12). Potential and current can be monitored using voltmeters, although this is also possible with the built-in instruments in the potentiostat. In most current applications of polarization techniques, the experiment is carried out using software that is either supplied by the instrument manufacturer or is available from other commercial sources. This method is a very convenient practice that allows storage of many experimental data that can later be analyzed with appropriate analysis software. Test or Working Electrode. The WE can be of various shapes, such as a cylinder or a flat disk. Often, the availability of a material as rod or sheet determines the shape of the electrode. When a cylinder is used, it has to be considered that edge effects can play a role due to different current distribution along the WE surface. The metallurgical properties may also be different for the sides and the bottom of the cylinder. Fewer problems occur with a flat disk, especially when care is taken that crevice corrosion does not occur. Not much attention is usually paid to the size of the electrode, with 1 cm2 (0.2 in.2) being a common test area. However, a much larger size is useful when very small currents have to be measured, while much smaller electrodes are useful when very large currents flow in the system. The error due to the ohmic drop between WE and RE is less when small electrodes are used. The WE is connected to an electrode holder to which the lead to the potentiostat is attached. Details of the construction of electrode holders appear in Ref 11 and 12 and are not repeated here. It is important that only glass, synthetic fluorine-containing resin, or other nonmetallic materials are immersed as part of the electrode holder. Figure 1 shows electrode holders with a cylinder electrode (Fig. 1a) or a sheet electrode (Fig. 1b).

Fig. 1 Test electrodes and electrode holders. (a) Cylindrical electrodes. (b) Flat disk or sheet electrodes. OD, outside diameter In neutral, aerated solutions, the cathodic reaction (reduction of oxygen) is usually under mass transport control. In this case, the electrochemical experiment should be carried out under controlled hydrodynamic conditions that can be achieved through the use of a rotating disk electrode (Fig. 2a) or a rotating cylinder electrode (RCE) (Fig. 2b). The RCE is preferred for corrosion studies because it provides very good current distribution. Polarization curves may be obtained at a constant rotation speed or as a function of rotation speed in studies where information concerning the details of the diffusion- controlled reduction reaction is required.

Fig. 2 Test electrodes. (a) Rotating disk electrodes. (b) Rotating cylinder electrodes Counter Electrode. The choice of the CE is dictated by the requirement that the CE should be of good conductivity and should not dissolve as a result of current flow. Platinum is used in most experiments. Because it is desirable to use CEs that are much larger than the WE in order to reduce the current density at the CE and avoid contamination by reaction products of the electrolyte, less expensive materials, such as graphite, stainless steel, or titanium alloys, are often used as CEs. For the same reason, the CE should be placed in a separate compartment with a glass frit to the main cell. However, this precautionary measure is often neglected. Reference Electrode. The saturated calomel electrode (SCE) is the most common RE used in studies of corrosion reactions, due to its stability and commercial availability. Care must be taken to avoid contamination of the test electrolyte by leakage of chloride ions from a SCE when tests are carried out in chloride ion-free test solutions. A mercury/mercurous sulfate electrode may be used when chloride ion contamination is not acceptable. Other REs are the Ag/AgCl and the Cu/ CuSO4 electrodes. A luggin capillary is used to minimize the ohmic drop between WE and RE (Ref 11, 12). A number of REs should be kept in the laboratory, with one RE being set aside as a calibration standard or the master RE, which is not used in actual potential tests. All other REs should be checked against this master RE before use by measuring the potential difference between the two electrodes, using a high-input impedance voltmeter. The potential difference should be less than 2 to 3 mV. Electrodes still showing a large potential difference after this procedure should be reconditioned or discarded. Test Cell. The design of the test cell can be quite complicated for studies in deaerated solutions with constant temperature control, or very simple in tests with oxygenated solutions at room temperature. Three-compartment cells are often used in the first case, while an open beaker often gives satisfactory results in the second case. For larger sheet electrodes, such as those used in the evaluation of polymer coatings or anodic films (Ref 16), the simple cell with the flat sample exposed horizontally, as shown in Fig. 3, has been used by many authors. The Avesta cell has the advantage that its cell design allows crevice corrosion, which might be a problem with most other cell designs, to be avoided during the test. Important points to consider in assembly of a test cell are the correct placement of the luggin capillary for minimization of the ohmic drop and the CEs for satisfactory current distribution. Deaeration of the test solution is carried out by bubbling nitrogen gas or argon through the test cell for a sufficient time.

Fig. 3 Test cell for flat sheet electrodes. CE, counter electrode; RE, reference electrode; WE, working electrode Potentiostat. A number of different potentiostats are available from different manufacturers. These instruments can be very sophisticated, allowing the recording of a number of electrochemical tests under computer control, or quite simple, with only the basic features in a low-cost model. Some newer instruments, using appropriate software, allow simultaneous recording of several experiments. The main requirements of a potentiostat have been discussed elsewhere in detail (Ref 11, 12). The available range of Eappl should be at least ±2 V, and the current capability should be ±1 A. A larger range of Eappl may be required for some tests, such as the pitting of titanium alloys. The maximum cell voltage should be approximately 60 V. Outputs should be provided for monitoring of potential and current. For the latter purpose, the current should be provided as a voltage versus ground, which is important for recording of the current on a strip chart or x-y recorder. The positive feedback or the current interrupter technique can be used for elimination of the ohmic drop, which can be significant in studies in low-conductivity media such as very dilute solutions and nonaqueous media (Ref 17, 18). The possibility of conducting the experiment under computer control is highly desirable, especially for more complicated tests such as the potentiodynamic tests described subsequently. Voltmeters. The values of potential and current usually are displayed on the front panel of the potentiostat. Often, only one of these parameters can be read at a given time. Therefore, it is sometimes convenient to observe both values on separate voltmeters, especially when meters with a digital display are used. If currenttime (I-t) or potential-current (E-I) curves are recorded, the use of digital voltmeters is useful to ensure that these curves are recorded correctly. If it is desired to determine the potential of an electrode directly with a voltmeter rather than from the potential output of the potentiostat, a high-input impedance voltmeter has to be used to prevent current flow and polarization of WE and RE due to the potential measurement. The potential can be recorded on a strip chart recorder using the output of such a voltmeter. In this case, the input impedance of the recorder is not important. However, direct connection of a recorder to the RE and WE should be avoided, because the input impedance of most recorders is quite low. Recording Devices. For tests in which the current, I, is monitored at a fixed value of Eappl as a function of time, t, a strip chart recorder is needed for the recording of the I-t curve. In this case, the current output, which is provided in the potentiostat as a voltage versus ground, is connected to the recorder, with the appropriate settings of the current range on the potentiostat and the voltage range on the recorder. For tests in which a number of potential steps are applied, increasing in the anodic or cathodic direction, a strip chart or an x-y recorder can be used for the recording of the I-t curve. In the latter case, the current output is connected to the y-axis, and the x-axis is set to a time scale that is appropriate for the time used for each potential step.

If a computerized system is used, recording devices are not needed. One of the advantages of the use of proper software is the fact that the electrochemical data are displayed on the screen as they are collected, which allows the researcher to stop an experiment that obviously is not running properly. Computer with Software Program and Plotter. It is convenient and time saving to use a software program to record polarization data. Most commercial software programs contain a number of typical electrochemical tests, such as potentiostatic, potentiodynamic, galvanostatic polarization, and other tests, such as the measurement of the galvanic current, the electrochemical impedance, or ECN. The user can input the particular parameters that are desired for use in the test. Usually, default values are included in the program, which allows the running of a test with only a few inputs provided by the user. For a potentiostatic test of type 1 (explained later in this article), the inputs are the applied potential, Eappl, and the time for which the potential is applied. Additional inputs depend on the particular software program to be used. The I-t curve is displayed on the computer screen with a current scale that has been selected by the program. After the test is completed, the I-t curve can be plotted after reformatting of the two axes into more convenient scales. In a galvanostatic test, a constant current, I, is applied for a certain time, and the E-t curve is monitored. The potentiostatic test of type 2 (detailed later) requires additional inputs, such as the starting potential and the final potential as well as step height and length. Sometimes, for cleaning or other purposes, the test electrode is held at a preconditioning potential for a certain period before the test is started. Figure 4 shows a typical arrangement for the application of electrochemical techniques in laboratory studies. The potentiostat connected to the test cell is controlled by a computer with suitable software; the experimental data can be printed out on the printer and/or displayed later, using software designed for analysis of the results of a given electrochemical test (Ref 11, 12).

Fig. 4 Experimental arrangement for application of electrochemical techniques in laboratory tests Thermostat. For studies that require carefully controlled temperatures, the test cell is placed in a constanttemperature bath, or a heating/cooling jacket is placed around the test cell.


11. B.J. Little, P.A. Wagner, and F. Mansfeld, Corrosion Testing Made Easy—Microbiologically Influenced Corrosion, B.C. Syrett, Ed., NACE, 1997 12. N.G. Thompson and J.H. Payer, Corrosion Testing Made Easy—dc Electrochemical Methods, B.C. Syrett, Ed., NACE, 1998 16. F. Mansfeld and M.W. Kendig, Corrosion, Vol 41, 1985, p 490 17. F. Mansfeld, Corrosion, Vol 32, 1976, p 143 18. F. Mansfeld, M.W. Kendig, and S. Tsai, Corros. Sci., Vol 22, 1982, p 455


No-Applied-Signal Tests In measurements of OCP or Ecorr, the galvanic current flowing between two dissimilar metals and ECN, no external electrochemical signal is applied. In one approach to the measurement of ECN data, the galvanic current flowing between two electrodes of the same material maintained at zero potential and the potential of this couple versus a stable RE are recorded simultaneously. Open-Circuit or Corrosion Potential. Measurement of Ecorr is the easiest electrochemical test to perform, but it provides the least amount of mechanistic information. Measurement of Ecorr requires a stable reference electrode, such as a SCE, a high-impedance voltmeter, and, in most cases, a suitable recording device. Corrosion potential can be determined also by observing the digital display on the potentiostat, monitored by recording the potential output value from the potentiostat on a strip chart recorder as a function of exposure time or, if a suitable software program is used, by plotting the potential- time curve stored on a disk on a printer after the test is finished. Open-circuit potential often increases with time for passive samples as passivity improves, and OCP often decreases when localized corrosion begins. The time dependence of Ecorr has been monitored by many groups investigating the effects of biofilm formation on localized corrosion of stainless steels in natural seawater (Ref 19). Most published data show a rapid ennoblement of Ecorr (i.e., a shift to more positive values) during the first days of exposure. The corrosion potential was monitored for Al 2024, mild steel, and cartridge brass during exposure to abiotic artificial seawater (AS) and to AS containing Shewanella algae or Shewanella ana (Ref 20). For Al 2024 (Fig. 5a) and brass (Fig. 5b), Ecorr decreased in the presence of bacteria, while for mild steel (Fig. 5c), an increase of Ecorr was observed. After exposure for 7 days to AS containing S. ana, kanamycin was added to kill the bacteria, and Ecorr changed to the values observed in the abiotic AS for Al 2024 and brass but not for mild steel (Fig. 5). Electrochemical impedance spectroscopy was used to determine the corrosion mechanism of these systems in more detail (Ref 20).

Fig. 5 Time dependence of corrosion potential (Ecorr) for (a) Al 2024, (b) brass, and (c) mild steel exposed to artificial seawater with or without bacteria. VNSS, Väätänen nine salts solution. K, kanamycin. Source: Ref 20 The main problem with the use of Ecorr measurements is overinterpretation. Based on information obtained with Ecorr alone, it is not possible to determine whether corrosion rates have increased or decreased. Some authors have interpreted shifts of Ecorr in the negative direction as a reduction of the rate of cathodic reaction or an increase of the rate of the anodic reactions. These are the two simplest possibilities that can be expected to lead to the observed result. However, without additional data, such as measurement of the polarization resistance (Rp), no valid mechanistic conclusions can be drawn. Dissimilar Metal Corrosion (Galvanic Corrosion). Short-circuit currents between dissimilar metals or metals exposed to electrolytes of different composition need to be measured by means of a ZRA, because the introduction of a milliammeter with an internal resistance, Ri, in a low-resistance circuit may alter the total current by producing a potential difference, IRi, between the two electrodes. The ZRA based on operational amplifiers, which was introduced in 1970 by Lauer and Mansfeld (Ref 21) for corrosion studies (Fig. 6), applies 0 mV potential difference between the two electrodes in the galvanic couple. The potential of the couple should be measured versus a RE in order to obtain additional information concerning the corrosion mechanism. A potentiostat can be used as a ZRA by setting the applied potential to 0 mV. The galvanic current is monitored as the current output of the potentiostat.

Fig. 6 Instrumental arrangement for continuous measurement of the galvanic current and potential. A1 and A2 represent operational amplifiers. Rm, measuring resistor; Vg, galvanic voltage; Vo, output voltage. Source: Ref 22 In using a ZRA based on operational amplifiers, one electrode is connected to the grounded input (Fig. 6). The galvanic current, Ig, needed to maintain 0 mV between the two electrodes is monitored as the output voltage Vo = IgRm, where Rm is the measuring resistor. The measuring resistor can be varied to maintain Vo in a convenient range, that is, 0 to 1 V. The potential Vg of the galvanic couple is measured with the voltage follower (A2 in Fig. 6), which determines the potential of the RE versus ground. Both Vo and Vg can be recorded on strip chart recorders or other devices. The advantage of the design in Fig. 6 is its low cost, which allows construction of several devices and monitoring of multiple test cells at the same time. Operational amplifiers can also be used to construct inexpensive and versatile potentiostats, as discussed elsewhere (Ref 22). The galvanic current is a measure of the increase of the corrosion current of the anode due to the coupling to the cathode and does not allow calculation of the corrosion rate of either electrode (Ref 23). Simultaneous measurements of the galvanic current and the potential of the coupled electrodes using a RE allow determination of the changes of the kinetics of the corrosion reactions with exposure time. It has to be recognized that the magnitude of the measured galvanic current depends on the area ratio of anode and cathode, Ac to Aa. If this ratio is not 1, then the measured galvanic current, Ig, has to be normalized to determine the and , which have different numerical values (Ref 24). galvanic current densities, Detailed studies have been conducted on the galvanic corrosion of aluminum alloys coupled to a number of materials (Ref 25, 26). Figure 7 shows the time dependence of the galvanic current, Ig, (Fig. 7a) and the potential, Eg, (Fig. 7b) of different galvanic couples versus a SCE (Ref 25). Galvanic current flowing between Al 2024 and a dissimilar metal decreased in the sequence copper > 304 stainless steel > Ti-6Al-4V, indicating that copper acted as an efficient cathode, while the titanium did not cause significant galvanic corrosion of Al 2024. For couples involving cadmium or zinc, Al 2024 was protected cathodically (Fig. 7a). Except for the couple involving zinc, Eg remained close to Ecorr of Al 2024 (Fig. 7b). Similar results were obtained for Al 7075 (Fig. 7). A ZRA has been used to evaluate the effect of dichromate additions on the galvanic corrosion of a brass-steel couple in aerated 20% NaCl (Ref 27).

Fig. 7 Time dependence of (a) galvanic current (Ig) and (b) galvanic potential (Eg) for Al 2024 and Al 7075 coupled to different dissimilar materials in aerated 3.5% NaCl. Source: Ref 25 Electrochemical Noise Analysis (ENA). Electrochemical noise data can be obtained as fluctuations of Ecorr; as fluctuations of the potential, E, at an applied current, I; or as fluctuations of I at an applied E. Recently, it has

become common practice—at least in laboratory studies—to measure potential and current fluctuations simultaneously, as shown in Fig. 8 (Ref 28, 29, 30, 31). In this approach, two electrodes of the same material are coupled through a ZRA. Current fluctuations are measured using the ZRA, while the potential fluctuations are measured with a high-impedance voltmeter between the two coupled electrodes and an RE, which could be a stable RE, such as an SCE, or a third electrode of the same material as the two test electrodes (Ref 28, 29). The latter approach is commonly used in corrosion monitoring. A review of different methods for collection of ECN data can be found in Ref 32.

Fig. 8 Experimental arrangement for simultaneous collection of potential and current fluctuations. Source: Ref 30 Simultaneous collection of potential and current ECN data allows ENA in the time and the frequency domains. Various methods of data analysis have been discussed in detail (Ref 30, 31, 33, 34, 35). The parameters commonly determined in these analyses are listed in Table 2. Figure 9 shows experimental potential, V, (Fig. 9a) and current, I, (Fig. 9b) fluctuations for mild steel exposed to aerated 0.5 N NaCl as well as the power spectral density (PSD) plots for these data (Fig. 9c, d) (Ref 36). The drift in the raw experimental data that causes erroneous PSD slopes was removed using a linear method (Ref 36). From the PSD plots for potential and current fluctuations, the spectral noise or noise impedance plot (Rsn) (Fig. 9e) can be determined as the ratio of the fast Fourier transform of potential and current fluctuations, or the square root of the ratio of the potential and current PSD plots. Good agreement was observed between the impedance spectra and the spectral noise plots (Ref 36). Analysis of ECN data in the time domain results in values of the mean potential, Ecoup, and the mean current, Icoup, of the coupled electrodes. In addition, the standard deviations σV and σI of the potential and current fluctuations, respectively, and the noise resistance Rn = σV/σI are obtained. The localization index LI = σI/Irms, where Irms is the root mean square value of the current fluctuations, can also be determined (Ref 32, 37). The suggestion that LI can be used to determine the prevailing corrosion mechanism should be viewed with great caution (Ref 35, 37). Table 2 Statistical analysis of ECN data Parameter Mean Standard deviation Root mean square

Formula

Localization index Noise resistance Skewness Kurtosis

Fig. 9 Electrochemical noise data after trend removal for mild steel exposed to 0.5 N NaCl. (a) Voltage time history. Voltage is measured against SCE. (b) Current time history. (c) Power spectral density (PSD) plot from voltage data. (d) Power Spectral density plot from current data. (e) Spectral noise impedance, Rsn, derived from PSD. Source Ref 36 It has been assumed often that Rn is equal to Rp; however, it has been shown for passive systems, such as stainless steel or titanium alloys exposed to neutral media, that Rn was several orders of magnitude lower than Rp (Ref 35, 38). Similar results were reported for polymer- coated steel. In fact, for epoxy-polyamide- coated steel, no relationship between Rn and any properties of the coated steel sample could be found (Ref 29, 30, 31). Moreover, Rn was found to depend on bandwidth Δf of the noise measurement (Ref 30). In general, Rn is close to Rp only in those cases where the impedance is independent of frequency within Δf (Ref 30). Skewness and kurtosis of the ECN data can be used for a more detailed analysis of the ECN data. High kurtosis values are considered to be indicative of localized corrosion. Drift of ECN data can affect the slopes of PSD plots (Ref 39). It has been shown how to correct derived parameters such as σV, σI, and Rn for the effects of drift (Ref 36). A description of the background and application of the ECN technique, including detailed descriptions of the measurement procedure and different approaches to the analysis of ECN data, is available in Ref 13.

References cited in this section 13. R. Cottis and S. Turgoose, Corrosion Testing Made Easy—Electrochemical Impedance and Noise, B.C. Syrett, Ed., NACE, 1999 19. F. Mansfeld and B. Little, Corros. Sci., Vol 32, 1991, p 247 20. A. Naguib and F. Mansfeld, Electrochim. Acta, Vol 37, 2002, p 2319

21. G. Lauer and F. Mansfeld, Corrosion, Vol 26, 1970, p 504 22. F. Mansfeld and R.V. Inman, Corrosion, Vol 35, 1979, p 21 23. F. Mansfeld, Corrosion, Vol 29, 1973, p 403 24. F. Mansfeld, Corrosion, Vol 27, 1971, p 436 25. F. Mansfeld, D.H. Hengstenberg, and J.V. Kenkel, Corrosion, Vol 30, 1974, p 343 26. F. Mansfeld and J.V. Kenkel, Corros. Sci., Vol 15, 1975, p 183, 239 27. D.A. Jones, Corrosion, Vol 40, 1984, p 181 28. F. Mansfeld, L.T. Han, and C.C. Lee, J. Electrochem. Soc., Vol 143, 1996, p L286 29. F. Mansfeld, L.T. Han, C.C. Lee, C. Chen, G. Zhang, and H. Xiao, Corros. Sci., Vol 39, 1997, p 255 30. F. Mansfeld, L.T. Han, C.C. Lee, and G. Zhang, Electrochim. Acta, Vol 43, 1998, p 2933 31. F. Mansfeld, H. Xiao, L.T. Han, and C.C. Lee, Prog. Org. Coatings, Vol 30, 1997, p 89 32. D.A. Eden, “Electrochemical Noise—The First Two Octaves,” Paper 386, Corrosion/ 98, NACE, 1998 33. U. Bertocci, C. Gabrielli, F. Huet, and M. Keddam, J. Electrochem. Soc., Vol 144, 1997, p 31 34. U. Bertocci, C. Gabrielli, F. Huet, M. Keddam, and P. Rousseau, J. Electrochem. Soc., Vol 144, 1997, p 37 35. F. Mansfeld, Z. Sun, E. Speckert, and C.H. Hsu, “Electrochemical Noise Analysis (ENA) for Active and Passive Systems,” Paper 00418, Corrosion/2000, NACE, 2000 36. F. Mansfeld, Z. Sun, C.H. Hsu, and A. Nagiub, Corros. Sci., Vol 43, 2001, p 341 37. F. Mansfeld and Z. Sun, Corrosion, Vol 55, 1999, p 915 38. C.C. Lee and F. Mansfeld, Corros. Sci., Vol 40, 1998, p 959 39. C. Gabrielli, F. Huet, and M. Keddam, Electrochim. Acta, Vol 31, 1986, p 1025


Small-Signal Polarization Tests

In measurements of the polarization resistance and electrochemical impedance spectra, only a small electrochemical signal is applied in order to ensure that the system under investigation remains linear. This requirement also ensures that these measurements are nondestructive and therefore can be repeated many times on the same system. Polarization-Resistance Method. Thorough reviews of background and application of the polarizationresistance technique for measuring corrosion currents are found in Ref 4 and 40. The polarization resistance (Rp) defined as: (Eq 1) is determined as the slope of a polarization curve at Ecorr, where i = 0. Usually, potentiodynamic polarization curves collected in close vicinity of Ecorr are used for this purpose. ASTM G 59, “Standard Practice for Conducting Potentiodynamic Polarization-Resistance Measurements,” deals with experimental procedures for determining Rp according to Eq 1. This method describes an experimental procedure for polarization-resistance measurements that allows an investigator to calibrate equipment and test methods. Figure 10 shows standard polarization curves for type 430 stainless steel in 1.0 N H2SO4, deaerated with hydrogen, at 30 °C (85 °F). Experimental polarization curves must fall within curves 2 and 3; otherwise, experimental problems have occurred.

Fig. 10 Standard potentiodynamic polarization curves in the vicinity of Ecorr for type 430 stainless steel in 1 N H2SO4, 30 °C (85 °F). Source: ASTM G 59 Applications of the Rp technique as a nondestructive method for monitoring corrosion rates have been discussed in detail in review articles (Ref 4, 40). The main advantage of the Rp technique over periodic weight-loss measurements is the possibility of continuously monitoring instantaneous corrosion rates of a metal exposed to a corrosive environment. Polarization resistance can be converted to icorr using the Stern-Geary equation: icorr = B/Rp

(Eq 2)

where B is a parameter calculated using the Tafel slopes: B = βaβc/2.3(βa + βc) where βa and βc are the anodic and cathodic Tafel slopes, respectively.

(Eq 3)

In order to obtain quantitative values of icorr from experimental Rp data, B has to be known for the same experimental conditions. For qualitative estimates, a constant value of B between 13 and 26 mV often is used. Corrosion rates can be obtained from icorr using Faraday's law. In determining Rp values from polarization curves, scan rate and ohmic drop effects have to be considered. Experimental polarization curves can change with increasing scan rate. In general, Rp is underestimated (and corrosion rates overestimated) when scan rates are too high. Polarization resistance is overestimated (and corrosion rates underestimated) when the uncompensated resistance (Rs) is not considered, because the experimental value of Rp (Rp,exp) contains a contribution from Rs: Rp,exp = Rp + Rs

(Eq 4)

The contribution from Rs can be eliminated during the measurement by using positive feedback or current interrupter techniques or by subtracting Rs from Rp,exp as calculated from analysis of the polarization curve. Uncompensated resistance can be determined by applying a constant current pulse, Iappl, and observing the potential response on an oscilloscope (Fig. 11). The initial sharp rise of the potential E = IapplRs due to ohmic drop is followed by a slower increase due to double-layer charging. Thus, Rs = E/Iappl.

Fig. 11 Potential pulse method for determining the uncompensated resistance, Rs The error e = Rs/Rp due to the uncompensated resistance can have similar values for a system with high corrosion rates in a solution of high conductivity and for a system with low corrosion rates in a solution of low conductivity. Therefore, ohmic drop effects can be important in all systems. Experimentally, it has been observed that the largest errors due to the uncompensated resistance occur for systems with high corrosion rates and low conductivity, such as steel in acid/ alcohol mixtures. The linear polarization technique is a variation of the polarization-resistance technique and forms the basis of commercial corrosion-rate monitors. In this approach, it is assumed that polarization curves are linear in the vicinity of Ecorr (Ecorr ± 10 mV). In this case, R′p is defined as: (Eq 5) where ΔE = Eappl - Ecorr, and i is the current density measured at the applied potential, E. Because ΔE is constant, icorr is proportional to the measured current density, i: icorr = k · i

(Eq 6)

where k is treated as a system constant. In commercial instruments, positive and negative values of ΔE are applied, and the average value of i is used to determine the corrosion rate. The assumption that the polarization curve is linear with ΔE has been discussed, and the limitation of the technique has to be considered in each application (Ref 4, 40). Several commercial devices are available, and automatic ohmic drop compensation is available in some instruments. Electrochemical impedance spectroscopy (EIS) data are recorded as a function of frequency of an applied alternating current (ac) signal at a fixed working point (E, i) of a polarization curve. In corrosion studies, this working point is often Ecorr (E = Ecorr, i = 0). Usually, a very large frequency range has to be investigated to obtain the complete impedance spectrum. In most corrosion studies, this frequency range extends from 100 kHz, the upper limit of commonly used frequency response analyzers (FRAs), to 1 mHz. Measurements at lower frequencies are very time consuming. Impedance data usually are determined with a three-electrode system, although it is also possible to use a two-electrode system in which both electrodes are of the same material. In this case, the EIS data are collected at E = 0 V. A potentiostat is used to apply the potential at

which the data are to be collected. The FRA is programmed to apply a series of sine waves of constant amplitude, small enough to remain in the linear potential region, and varying the frequency. Impedance data are determined by the FRA at each frequency and stored in its memory. Because a very large number of data points have to be collected, displayed, and analyzed, it is essential to use adequate software for these purposes. Such software is available from all major manufacturers of FRAs. Information obtained with EIS differs from that determined with the other techniques previously described, because the corrosion system is analyzed at a fixed potential (or current). Properties of the system at this potential can then be determined through analysis of the frequency dependence of the impedance. One of the advantages of EIS is that only very small signals are applied, making it a nondestructive technique. In many cases, EIS data have shown that existing models for corrosion systems based on results of studies with directcurrent (dc) techniques were too simplistic. Experimental impedance spectra can be displayed using linear coordinates in complex plane plots in which the negative value of the imaginary impedance (-Z″) is plotted versus the real part (Z′). For a simple circuit consisting of a parallel combination of resistance, Rp, and capacitance, Cdl, with Rs in series (Fig. 12a), such a plot results in a semicircle (Fig. 12b). This approach neglects the fact that the impedance often changes over many orders of magnitude, and that the frequency dependence of the impedance cannot be recognized in complex plane plots. The preferred format for display of EIS data is therefore the Bode plot (Fig. 12c, d), in which log |Z| and the phase angle (Φ) are plotted versus the logarithm of the frequency (f) of the applied signal, with |Z| being the modulus of the impedance. In Fig. 12, the curves for α = 1 refer to ideal capacitive behavior, while the curves for α = 0.9 are typical for the often-observed deviations from ideal behavior. The constant phase element concept is often used in such cases. ASTM G 106, “Standard Practice for Verification of Algorithm and Equipment for Electrochemical Impedance Measurements,” describes an experimental procedure that may be used to check the instrumentation and technique for collection and presentation of EIS data. Reference impedance plots in both complex plane and Bode format are included for a dummy cell and for stainless steel type 430 in a solution containing 0.005 M sulfuric acid (H2SO4) and 0.495 M sodium sulfate (Na2SO4).

Fig. 12 Representations of a corroding electrode. (a) Equivalent circuit model. Rs, 10 Ω; Rp, 1000 Ω; Cdl, 20 μF. (b) The complex plane or Nyquist plot of the equivalent circuit for α = 1 and α = 0.9. Note the intercepts with the real axis occur at Rs and Rs + Rp. (c) Bode plot of the base 10 log of the magnitude of the impedance against the log of the frequency. The range of the impedance is between Rs and Rs + Rp. (d) Bode plot of the negative phase angle versus the log of the frequency In order to perform a quantitative analysis of EIS data, it is necessary to use an appropriate equivalent circuit that describes accurately the physical properties of the system under investigation. Equivalent circuits for polymer-coated metals, anodized aluminum alloys, and localized corrosion of aluminum alloys have been described in the literature (Ref 41). The ANALEIS software contains modules that were designed for the analysis of EIS data for these systems (Ref 41). The most successful application of EIS has probably been in the analysis of the properties of protective polymer coatings and their changes during exposure to corrosive environments (Ref 42, 43). Impedance spectra for polymer-coated metals exposed to corrosive media can be analyzed with the software module COATFIT, which is based on the equivalent circuit shown in Fig. 13(a). In this equivalent circuit, Cc is the capacitance of the coating, which depends on its dielectric constant, ε, and its thickness, d: Cc = εεoA/d

(Eq 7)

where εo is the dielectric constant of vacuum, and A is the exposed area. The variable Rpo has been called the pore resistance, because it is considered to be due to the formation of ionically conducting paths in the polymer (Ref 44, 45). The variable Rp is the polarization resistance of the area at the polymer coating/metal interface at which corrosion occurs, and Cdl is the corresponding capacitance. An increase of Cc with time can be related to water uptake by the coating (Ref 43). The rate at which Rpo decreases with increasing exposure time has been used to rank the efficiency of different pretreatments for steel in providing corrosion protection by a polybutadiene coating (Ref 44, 45). The changes of Cdl can be used to estimate the corroding delaminated area, Acorr, (Ref 46) and to calculate the specific polarization resistance from which an estimate of the corrosion rate of the delaminated area can be made. Figure 13(b) shows a comparison of experimental and fitted impedance spectra for polymer-coated Mg AZ31 that had been exposed to 0.5 N NaCl. Excellent agreement of the experimental and fitted data was observed. The use of EIS for lifetime prediction of organic coatings on steel is described in Ref 47. Damage functions based on the experimental values of Acorr have been determined and can be used for lifetime prediction of polymer-coated steel exposed to natural seawater for several years (Ref 48).

Fig. 13 (a) Equivalent circuit for polymer-coated metals exposed to a corrosive environment. (b) Comparison of an experimental and fitted impedance spectra for polymer-coated Mg AZ31 exposed to 0.5 N NaCl. (b) Cc, 5.129 pF; Cdl, 99.69 pF; Rpo, 357.1 Ω; Rs, 3Ω; Rp, 102.1 kΩ. Source: Ref 41 Impedance spectra for anodized aluminum alloys can be analyzed with the module ANODAL (Ref 41). The analysis is based on the equivalent circuit in Fig. 14(a), where Cpo is the capacitance of the outer porous layer, Rpo is the resistance of the sealed pores, and Cb and Rb are the capacitance and resistance of the inner barrier layer, respectively (Ref 16, 49, 50, 51). Analysis of impedance spectra for a large number of anodized and sealed aluminum alloys has shown that the spectra fall into a narrow scatter band if the same anodizing and sealing procedures are followed. Electrochemical impedance spectroscopy has been recommended therefore as

a quality-control test for anodized and sealed aluminum alloys (Ref 16). It should be noted that for unsealed or dichromate sealed samples, Rpo is very low. Therefore, Cpo does not appear in the measured spectrum in these two cases (Ref 51).

Fig. 14 (a) Equivalent circuit for anodized aluminum alloys. (b) Comparison of experimental and fitted spectra for Al 6061 (sulfuric acid anodized and hot water sealed) exposed to 0.5 N NaCl. Cpo, 56.87 pF; Cb, 0.37 μF; Rpo, 28.83 kΩ; Rb, 26.62 MΩ; Rs, 1Ω. Source: Ref 41 Figure 14(b) shows the result of a fit of experimental data for sulfuric-acid-anodized and hot-water-sealed Al 6061 that had been exposed to 0.5 N NaCl. Good agreement between the measured and fitted spectra was observed at the highest and lowest frequencies. However, significant deviations occurred in the frequency range between 1000 and 0.1 Hz. For a modified equivalent circuit in which Rpo was replaced by a transmission line impedance Zpo = K(jω)n (Fig. 15a), excellent agreement between the experimental and fitted data was obtained (Fig. 15b). From the fitted values of Cb and Cpo, the thickness of the barrier and the porous layer can be obtained. For the fit parameters in Fig. 15(b), a barrier layer thickness of 250 Å was calculated. The calculated thickness of 21 μm for the porous layer agreed very well with that obtained by observation in the scanning electron microscopy of the cross section of the anodized sample (Ref 41).

Fig. 15 (a) Modified equivalent circuit for anodized aluminum alloys. (b) Comparison of experimental and fitted spectra for Al 6061 (sulfuric acid anodized and hot water sealed) exposed to 0.5 N NaCl. Source: Ref 41 The experimental impedance spectra for aluminum alloys exposed to solutions containing chloride ions show characteristic changes when pits initiate and grow (Ref 52, 53, 54, 55). The spectra can be explained by the equivalent circuit in Fig. 16, where Rp and Cp are the polarization resistance and capacitance, respectively, of the passive surface; Rpit and Cpit are the corresponding values for the growing pits; W = (K/F)(jω)n is a transmission line impedance; and F is the area fraction at which pits have initiated. Figure 17 gives a comparison of experimental and fitted impedance spectra for Al 6061 after 24 h exposure to 0.5 N NaCl. As explained elsewhere (Ref 56), it is possible to obtain F from the time dependence of Cpit, which allows determination of the pitted area, Apit, which can be used to normalize Rpit and obtain the specific pit polarization

resistance, As a result of this procedure, it is possible to determine relative pit growth rates, the time dependence of which can be expressed as: (Eq 8)

Fig. 16 Equivalent circuit for the impedance of aluminum alloys for which pits have been initiated. Source: Ref 41

Fig. 17 Comparison of experimental (curve 1) and fitted (curve 2) impedance spectra for Al 6061 after 24 h exposure to 0.5 N NaCl. Source: Ref 41 The time dependence of log 18 agrees with the time law in Eq 8.

for Al 7075 that had been passivated in 1000 ppm CeCl3 (Ref 55) in Fig.

Fig. 18 Time dependence of Source: Ref 41

for Al 7075 (passivated in CeCl3) during exposure to 0.5 N NaCl.

Electrochemical impedance spectroscopy data have often been used only to determine Rp, which is defined as the limit of the real part of the impedance for f = 0 (Ref 41): (Eq 9) In this case, all other important information contained in the impedance spectrum is ignored. Nevertheless, the nondestructive nature of EIS makes it an ideal tool for monitoring the corrosion behavior over extended time periods. In Fig. 19(a), Rp values obtained for Al 2024 exposed to artificial seawater (AS) containing S. algae or S. ana are plotted as a function of exposure time, while the corresponding results for brass and mild steel are plotted in Fig. 19(b) and (c), respectively (Ref 20). For Al 2024 exposed to sterile AS, pitting was observed, and therefore,

obtained by fitting of the experimental spectra to the equivalent circuit in Fig. 16 is shown in

Fig. 19(a). Similar very high values were obtained for Al 2024 in the presence of S. algae or S. ana, indicative of passive behavior. When kanamycin was added to the AS containing S. ana after 7 days, Rp did not change immediately (Fig. 19a). After 15 days, pitting was observed. These changes of Rp are different from those of Ecorr in Fig. 5(a), where Ecorr increased very shortly after addition of kanamycin. For brass, S. algae produced only a small increase of for Al 2024 (Fig. 19a),

, while a much larger effect was found for S. ana (Fig. 19b). As observed

did not change for several days after addition of kanamycin (Fig. 19b), while Ecorr

changed almost immediately (Fig. 5b). The changes of

due to the addition of the two different bacteria were

quite similar for mild steel, and, as observed for Al 2024 and brass, in sterile AS for several days (Fig. 19c).

did not decrease to the levels observed

Fig. 19 Time dependence of Rp for (a) Al 2024, (b) brass, and (c) mild steel during exposure to sterile artificial seawater and artificial seawater containing S. algae or S. ana. VNSS, Väätänen nine salts solution. K, kanamycin. Source: Ref 20

References cited in this section 4. F. Mansfeld, The Polarization Resistance Technique for Measuring Corrosion Currents, Advancements in Corrosion Science and Technology, Vol 6, M.G. Fontana and R.W. Staehle, Ed., Plenum Press, 1976, p 163 16. F. Mansfeld and M.W. Kendig, Corrosion, Vol 41, 1985, p 490 20. A. Naguib and F. Mansfeld, Electrochim. Acta, Vol 37, 2002, p 2319 40. J.R. Scully, Corrosion, Vol 56, 2000, p 199 41. F. Mansfeld, H. Shih, H. Greene, and C.H. Tsai, Analysis of EIS Data for Common Corrosion Processes, STP 1188, ASTM, 1993, p 37 42. F. Mansfeld, J. Appl. Electrochem., Vol 25, 1995, p 187 43. F. Mansfeld and M. Kendig, Electrochemical Impedance Tests for Protective Coatings, STP 866, ASTM, 1985, p 122 44. F. Mansfeld, M. Kendig, and S.Tsai, Corrosion, Vol 38, 1982, p 478 45. M. Kendig, F. Mansfeld, and S. Tsai, Corros. Sci., Vol 23, 1983, p 317 46. J. Titz, G.H. Wagner, H. Spaehn, M. Ebert, K. Juettner, and W.J. Lorenz, Corrosion, Vol 46, 1990, p 221 47. M. Kendig and F. Mansfeld, Mat. Res. Soc. Symp. Proc., Vol 115, 1988, p 193 48. F. Mansfeld, C.C. Lee, L.T. Han, B.J. Little, P. Wagner, R. Ray, and J. Jones-Meehan, “The Impact of MIC on Protective Polymer Coatings,” Final Report, Contract N00014- 94-1-0026, The Office of Naval Research, Aug 1998 49. J. Hitzig, K. Juettner, W.J. Lorenz, and W. Paatsch, Corros. Sci., Vol 24, 1984, p 945 50. J. Hitzig, K. Juettner, W.J. Lorenz, and W. Paatsch, J. Electrochem. Soc., Vol 133, 1986, p 133 51. F. Mansfeld and M.W. Kendig, J. Electrochem. Soc., Vol 135, 1988, p 828 52. F. Mansfeld, S. Lin, S. Kim, and H. Shih, Corros. Sci., Vol 27, 1987, p 997 53. F. Mansfeld and H.Shih, J. Electrochem. Soc., Vol 135, 1987, p 332 54. F. Mansfeld, S. Lin, S. Kim, and H. Shih, Electrochim. Acta, Vol 34, 1989, p 1123 55. F. Mansfeld, S. Lin, S. Kim, and H. Shih, J. Electrochem. Soc., Vol 137, 1990, p 78 56. F. Mansfeld, H. Shih, H. Greene, and C.H. Tsai, Detection and Monitoring of Localized Corrosion with EIS, STP 1188, ASTM, 1993, p 297


Large-Signal Polarization Tests In large-signal polarization tests, the applied potential is varied over a wide range, using either a series of applied potential steps or a change of potential at a constant rate. It is also possible to use a series of current steps or a constantly changing current; however, these techniques are less popular. Because polarization over a wide potential range might cause irreversible changes of the sample surface properties, it is not possible, in most cases, to repeat a large-signal polarization test on the same sample. Potentiostatic Techniques. There are two types of potentiostatic techniques: • •

Type 1: recording of the current as a function of time at a constant applied potential, Eappl Type 2: recording of the steady-state current as a function of Eappl

The data obtained in the type 1 test provide information about the time dependence of the rate of an electrochemical reaction at a given potential. The data collected with the type 2 test can be used to construct a polarization curve, which usually is plotted as Eappl versus the logarithm of the measured current density, i. Polarization curves also can be obtained by the potentiodynamic technique. Because the instrumentation and experimental procedures are quite similar for both techniques (Ref 11, 12), much of the discussion presented in this section is also valid for the following section. ASTM G 5, “Standard Reference Test Method for Making Potentiostatic and Potentiodynamic Anodic Polarization Measurements,” is a classic document that has been used for many years in many countries. This test method describes experimental procedures for checking one's experimental technique and instrumentation. If followed, using ASTM-provided type 430 stainless steel in 1.0 N H2SO4, the method provides repeatable polarization curves that reproduce the curves shown in ASTM G 5. If discrepancies are found, the experiment has not been performed correctly, there might be problems with the test cell, and/ or the equipment might not function properly. A complete polarization curve consists of a cathodic part and an anodic part. The cathodic portion of the polarization curve contains information concerning the kinetics of the reduction reaction(s) occurring for a particular system. Depending on the solution composition, a mass- transport-controlled region can be reached at more negative potentials, at which the reaction rate (the measured current density) depends only on the solution composition and the hydrodynamic conditions. The particular features of the anodic part of the polarization curve depend strongly on the metal-electrolyte system. Usually, a charge-transfer-controlled region occurs at potentials close to Ecorr. So-called passive metals show an active-passive transition followed by a passive region and the region of oxygen evolution at the highest applied potentials. For those metals, which are susceptible to localized corrosion, a large increase of the current occurs in the passive region when the pitting potential, Epit, has been exceeded. ASTM G 3, “Standard Practice for Conventions Applicable to Electrochemical Measurements in Corrosion Testing,” provides conventions for reporting and displaying electrochemical data and contains theoretical and experimental polarization curves. In general, only qualitative information, such as changes in the polarization characteristics in the passive region or changes of Epit due to the presence of inhibitors or alloying, is extracted from a full polarization curve. However, it is possible to determine from such curves quantitative information that is related to the kinetics of the corrosion reactions and the corrosion rate. The relationship between the applied potential, Eappl, and the measured current density, i, in the Tafel region is given by (Ref 2, 3): Eappl = Ecorr + βa log (i/icorr) for the anodic polarization curve and:

(Eq 10)

Eappl = Ecorr - βc log (i/icorr)

(Eq 11)

for the cathodic polarization curve. In Eq 10 and 11, the anodic Tafel slope, βa, and the cathodic Tafel slope, βc, are positive numbers. Plots of Eappl versus log i are called Tafel plots. The corrosion current density, icorr, can be determined according to Eq 10 and 11 by extrapolation of the Tafel lines to Eappl = Ecorr, where i = icorr (Fig. 20). This method of obtaining corrosion current density is called the Tafel extrapolation method. It is also possible to determine icorr from either the anodic or the cathodic Tafel plot or from the intersection of both plots. It is advantageous to use computer software to record polarization curves and analyze the experimental data in terms of parameters such as βa, βc, and icorr. From the numerical values of βa and βc, conclusions concerning the rate- determining step in the reaction mechanism can be made (Ref 2, 3, 4). From icorr, the corrosion rate can be calculated using Faraday's law. ASTM G 102, “Standard Practice for Calculation of Corrosion Rates and Related Information from Electrochemical Measurements,” provides guidance in converting the results of electrochemical measurements to rates of uniform corrosion.

Fig. 20 Theoretical Tafel plots illustrating the Tafel extrapolation method. SCE, saturated calomel electrode; SHE, standard hydrogen electrode

An equivalent circuit for a three-electrode cell is shown in Fig. 21, where Rp is the polarization resistance of the test electrode; C is its capacitance; Rs is the resistance of the electrolyte between the surface of the test electrode and the tip of the Luggin capillary, which houses the reference electrode; and Rc is the resistance of the electrolyte between the test electrode and the counter electrode. The applied potential, Eappl, appears across Rp + Rs and therefore always includes an ohmic drop, IRs, where I is the current flowing as a result of polarization. This ohmic drop is an error in the measurement, because the potential Eappl set at the potentiostat differs from the true electrode potential E by IRs. This fact has to be recognized in all polarization experiments. In many cases, where electrolytes with reasonable conductivity are used and IRs is small compared to Eappl, this error can be neglected. In other cases, it can be eliminated by the positive feedback or current interrupter techniques (Ref 17, 18) that are available in most modern potentiostats.

Fig. 21 Dummy cell for testing of potentiostat Experimental Procedure for Potentiostatic Test Type 1. The experimental procedures for the two types of potentiostatic tests are described using the same types of electrodes and test cells. They differ insofar as type 2 involves the application of a series of potential steps. This requires changes of Eappl by hand, periodic adjustment of the sensitivity of the current-measuring device in the potentiostat, and sometimes, adjustment of the current scale of the recorder also. The procedure is greatly simplified if appropriate software is used and the experiment is carried out under computer control. When the test cell has been assembled, the electrodes can be connected to the potentiostat. In most cases, Ecorr is monitored before starting the potentiostatic test. When Ecorr has reached a reasonable stable value, the potential can be applied and the corresponding current measured. Often, a series of potential steps is applied to the same test electrode. This is the case in a test in which the pitting potential, Epit, is determined as the potential at which the anodic current keeps increasing rather than decreasing after application of the potential step, as it would in the case for E < Epit. When the experiment is performed under computer control, the control parameters, such as the value of Eappl and the length of the potential step, have to be entered. Additional parameters related to the current measurement, such as the number of data points per second, might have to be entered. The experiment is performed then as programmed, and the current-time plot is displayed on the computer screen. After reformatting of the data, a printout of the I-t curve is obtained. Experimental Procedure for Potentiostatic Test Type 2. The test is started by applying the first potential, which might be 10 to 50 mV more positive than Ecorr if an anodic polarization curve is measured. After the current, which is recorded continuously throughout the test on a strip-chart recorder, has reached a steady-state value or after a predetermined time has elapsed, Eappl is increased to the next value, and the current is recorded. As increasing values of Eappl are used, it may become necessary to switch to less sensitive current scales. After the current at the final value of Eappl has been recorded, the polarization curve (Eappl versus log i) can be plotted using the final values of the current, I, at each potential and converting it into current density, i. Figure 22 shows the typical standard potentiostatic anodic polarization plot for type 430 stainless steel in 1.0 N H2SO4 at 30 °C (85 °F). Polarization curves performed according to ASTM G 5 with the standard ASTM sample have to fall within the scatter band shown in Fig. 22. Slightly different curves have been obtained with a new lot of samples introduced in recent years. Deviations from these standard plots point to deficiencies in the experimental technique, problems with the test cell, the equipment, and/or the operator.

Fig. 22 Standard potentiostatic polarization curve for type 430 stainless steel in 1.0 N H2SO4, 30 °C (85 °F). Step rate, 50 mV for 5 min (each specimen). Source: ASTM G 5 Potentiodynamic Techniques. Potentiodynamic polarization curves can be obtained as a single sweep, that is, Eappl is increased at a constant scan rate, s = dEappl/dt, between a starting potential, Es, which often is Ecorr or possibly below, and a final potential, Ef. Sometimes the scan is reversed when Ef or a preset maximum current density is reached, and stopped when the previous Ecorr or a new value of Ecorr is reached. Great care has to be taken in choosing an appropriate value of the scan rate (Ref 57). In general, less distortion of the true polarization curve occurs at the lowest scan rates. In ASTM G 5, a scan rate of 600 mV/h is recommended. Figure 23 shows the effects of varying scan rate on the shape of the anodic polarization curves for type 430 stainless steel in 1.0 N H2SO4 (Ref 14). Scan rates were increased from curve 1 to curve 4.

Fig. 23 Effect of scan rate on anodic polarization curve for type 430 stainless steel in 1.0 N H2SO4, 30 °C (85 °F). Scan rates, mV/s: 1) 0.03; 2) 0.17; 3) 0.34; 4) 1.00. Source: Ref 14 Single Sweeps. Anodic and cathodic polarization curves are usually determined at constant scan rate in a single sweep between two preset potentials, Es and Ef. ASTM G 3 gives examples of polarization curves determined in this manner. Figure 24 shows the typical standard potentiodynamic anodic polarization plot for type 430 stainless steel in 1.0 N H2SO4 at 30 °C (85 °F). Anodic potentiodynamic polarization curves performed according to ASTM G 5 with the standard ASTM sample have to fall within the scatterband shown in Fig. 24. Slightly different curves have been obtained with a new lot of samples introduced in recent years. Deviations from these standard plots point to deficiencies in the experimental technique or problems with the test cell, the equipment, and/or the operator.

Fig. 24 Standard potentiodynamic polarization curve for type 430 stainless steel in 1.0 N H2SO4, 30 °C (85 °F). 0.6 V/h. Source: ASTM G 5 The results obtained by Ecorr measurements and EIS suggest that the mechanism of microbiologically influenced corrosion inhibition (MICI) observed in the presence of Shewanella algae or ana is different for Al 2024 (Fig. 5a and 19a) and brass (Fig. 5b and 19b) on the one hand and mild steel (Fig. 5c and 19c) on the other. For Al 2024 and brass, the reduction of corrosion rates was accompanied by a shift of Ecorr to more negative values, suggesting that MICI is due to a decrease of the rate of oxygen reduction in the presence of the biofilm. The increase of Ecorr with a decrease of corrosion rates observed for mild steel can be explained only by a simultaneous decrease of the rates of the anodic and cathodic reactions. The former is assumed to be due to an increase of the Fe2+ concentration in the solution due to reduction of Fe3+ by Shewanella, leading to an increase of reversible potential, E0, of the Fe/Fe2+ reaction, as discussed in Ref 58. This mechanism is shown schematically in Fig. 25, where it is assumed that the anodic reaction is under charge transfer control, while the cathodic reaction is under charge transfer control at low polarization and under mass- transport control at higher polarization. In the presence of the biofilm, the limiting current density, iL, for oxygen reduction for Al 2024 and brass is decreased, but the rate of the anodic reaction remains unchanged. As a consequence, Ecorr decreases from point 1 in Fig. 25 to point 2, and iL = icorr decreases to i′L. For mild steel, the rate of both reactions is decreased, and Ecorr changes from point 1 to point 3, while iL decreases to i′L (Fig. 25).

Fig. 25 Schematic polarization curves for a charge-transfer-controlled anodic reaction and a masstransport-controlled cathodic reaction The assumptions leading to the proposed mechanism have been confirmed by recording of potentiodynamic polarization curves after exposure to Väätänen nine salts solution (VNSS, pH 7.5) containing Shewanella algae for 7 days (Ref 20). For Al 2024 (Fig. 26) and brass (Fig. 27), a significant reduction of the rate of the cathodic reaction was observed, while for mild steel (Fig. 28), the rate of both reactions was reduced.

Fig. 26 Potentiodynamic polarization curves for Al 2024 exposed for 7 days to artificial seawater containing S. algae. VNSS, Väätänen nine salts solution. Source: Ref 20

Fig. 27 Potentiodynamic polarization curves for brass exposed for 7 days to artificial seawater containing S. algae. VNSS, Väätänen nine salts solution. Source: Ref 20

Fig. 28 Potentiodynamic polarization curves for mild steel exposed for 7 days to artificial seawater containing S. algae. Source: Ref 20 Pitting Scans. Pitting scans are carried out to determine Epit and sometimes also the protection potential, Eprot. ASTM G 61, “Standard Test Method for Conducting Potentiodynamic Polarization Measurements for Localized Corrosion Susceptibility of Iron-, Nickel-, or Cobalt-Base Alloys,” gives a procedure for conducting cyclic potentiodynamic polarization measurements to determine relative susceptibility to localized corrosion in the form of pitting or crevice corrosion. An indication of the susceptibility to localized corrosion is given by the potential, Epit, at which the anodic current increases rapidly from the low values in the passive region. The more noble Epit with respect to Ecorr, the less susceptible is the alloy to initiation of localized corrosion (Ref 3, 11, 12). Figure 29 shows representative cyclic potentiodynamic polarization curves for type 304 stainless steel, which is susceptible to pitting and has a value of Epit close to 0 mV versus SCE, and alloy C-276, which is immune to localized corrosion. The area under the curve for the forward and reverse scan for a susceptible alloy is a qualitative measure of corrosion damage due to pit growth. The difference between pitting and protection potentials is also used for this purpose.

Fig. 29 Cyclic potentiodynamic polarization curves (pitting scans) for type 304 stainless steel and alloy C-276 in 3.54 wt% NaCl. Source: ASTM G 61 Evaluation of Alloy Sensitization. The electrochemical potentiokinetic reactivation (EPR) method, based on the original work found in Ref 59, has been developed for laboratory and field determination of the degree of sensitization of stainless steels (Ref 60, 61). The method is described in ASTM G 108, “Standard Test Method for Electrochemical Reactivation (EPR) for Detecting Sensitization of AISI Type 304 and 304L Stainless Steel.” It also has been used successfully to evaluate sensitization of other stainless steels and nickel alloys. The EPR test is carried out using a reactivation potentiodynamic sweep from the passive region back to the active region, stopping at Ecorr. The sweep is carried out in H2SO4 containing potassium thiocyanate (KSCN), which serves to activate the chromiumdepleted grain boundaries. The current peak observed during the reactivation scan increases with the degree of sensitization. The charge Q passed is a measure of the degree of sensitization and can be used to determine a sensitization number, Pa, after normalization with the grain- boundary area. A modification of the EPR method is the double-loop electrokinetic repassivation test that involves a potentiodynamic scan from Ecorr to the passive range, followed immediately by reverse polarization back to Ecorr (Ref 62, 63). In this method, the degree of sensitization is determined by the ratio Ir/Ia of the maximum current generated in the reverse scan compared to that in the initial anodic scan. Figure 30 illustrates the principle of these two methods.

Fig. 30 Schematic potentiodynamic polarization curves for two methods for determination of the degree of sensitization. (a) Clarke et al method (Ref 61). (b) Akashi et al method (Ref 62) Polarization curves obtained in the vicinity of Ecorr can be analyzed by fitting of the experimental data to the equation that relates potential, E, and current density, i (Ref 2, 3): (Eq 12) The polarization curves usually are obtained in the potentiodynamic mode. The results of the fit consist of experimental values of icorr, βa, and βc. Several software programs have been described for carrying out this fit (Ref 64). Most potentiostat manufacturers supply software for this procedure. Because only a small polarization range /E - Ecorr/ is applied (usually ±30 mV or less), this approach to obtain corrosion rates and Tafel slopes in the non-Tafel region is nondestructive and can be applied repeatedly without the danger of changing the properties of the surface under investigation. It is noted that the experimental approach is similar to that used in the determination of the polarization resistance, Rp. However, the kinetic data are obtained directly from the analysis of the measured polarization curve. Galvanostatic and galvanodynamic techniques are used less commonly than the potentiostatic and potentiodynamic techniques described previously. In galvanostatic techniques, a constant current is applied, and

the potential is monitored as a function of time. It is also possible to record an entire polarization curve in the galvanostatic mode. A limitation of this approach for passive metals is that the passive region cannot be investigated. ASTM G 100, “Standard Test Method for Conducting Cyclic Galvanostaircase Polarization,” describes a method for determining Epit for aluminum alloys, using a variation of the galvanostatic technique.

References cited in this section 2. H. Kaesche, The Corrosion of Metals, NACE, 1985 3. D.A. Jones, Principles and Prevention of Corrosion, MacMillan Publishing Co., 1992 4. F. Mansfeld, The Polarization Resistance Technique for Measuring Corrosion Currents, Advancements in Corrosion Science and Technology, Vol 6, M.G. Fontana and R.W. Staehle, Ed., Plenum Press, 1976, p 163 11. B.J. Little, P.A. Wagner, and F. Mansfeld, Corrosion Testing Made Easy—Microbiologically Influenced Corrosion, B.C. Syrett, Ed., NACE, 1997 12. N.G. Thompson and J.H. Payer, Corrosion Testing Made Easy—dc Electrochemical Methods, B.C. Syrett, Ed., NACE, 1998 14. F. Mansfeld, Corrosion, Vol 44, 1988, p 856 17. F. Mansfeld, Corrosion, Vol 32, 1976, p 143 18. F. Mansfeld, M.W. Kendig, and S. Tsai, Corros. Sci., Vol 22, 1982, p 455 20. A. Naguib and F. Mansfeld, Electrochim. Acta, Vol 37, 2002, p 2319 57. F. Mansfeld and M. Kendig, Corrosion, Vol 37, 1981, p 545 58. M. Dubiel, C.H. Hsu, C.C. Chien, F. Mansfeld, and D.K. Newman, Microbial Iron Respiration Can Protect Steel from Corrosion, Appl. Environ. Microbiol., submitted for publication 59. V. Cihal, A. Desestret, M. Froment, and G.H. Wagner, Proc. Conf. European Federation of Corrosion (Paris), European Federation of Corrosion, 1973, p 249 60. K. Osozawa, K. Bohnenkamp, and H.J. Engell, Corros. Sci., Vol 6, 1966, p 421 61. W.L. Clarke, R.L. Cowan, and W.L. Walker, in Intergranular Corrosion of Stainless Alloys, STP 656, ASTM, 1978, p 99 62. M. Akashi, T. Kawamoto, and F. Umemura, Corros. Eng. (Jpn.), Vol 29, 1980, p 163 63. A.P. Majidi and M.A. Streicher, Corrosion, Vol 42, 1984, p 584 64. H. Shih and F. Mansfeld, STP 1154, ASTM, 1992, p 174


Scanning Electrode Techniques The local distribution of anodic and cathodic sites can be evaluated using a number of electrochemical tests, such as the scanning reference electrode (SRET), the scanning vibrating electrode (SVET), local electrochemical impedance spectroscopy (LEIS), and the scanning Kelvin probe (SKP). While SRET and SVET allow measurements of the local corrosion currents, SKP provides a record of the surface potential distribution. The application of these techniques in electrochemistry and corrosion technology has recently been reviewed (Ref 65). A review of the early applications of scanning electrode techniques is found in Ref 8. The SRET has been used mainly for investigation of local corrosion processes, such as pitting and crevice corrosion (Ref 8). The application of the SRET technique to localized corrosion phenomena is reviewed in Ref 66. More recently, SRET has also been used for the evaluation of delamination of organic coatings (Ref 67). The SRET technique is based on a microprobe that is capable of detecting local variations of the electrochemical properties of a metal surface with high lateral resolution. An optimized SRET has been used to evaluate the performance of several clear coatings, with good correlation between SRET signals and permeability as well as wet adhesion of these coatings (Ref 68). The SVET offers a rapid method for monitoring the progress of corrosion of many metals and alloys exposed to corrosive solutions. The SVET technique has been shown to be a powerful tool for evaluating the efficiency of a variety of potential corrosion inhibitors as well as the determination of the current distribution in the vicinity of defects on polymer-coated metal surfaces (Ref 65, 69). It has also been used for determining the heterogeneous characteristics of consortia of bacteria and their possible influences on metallic substrates leading to microbiologically influenced corrosion (Ref 70). The SKP. While SRET and SVET have to be used in immersion, the SKP is normally operated in (moist) air (Ref 71, 72). The SKP measures the Volta potential difference between a vibrating microreference electrode and the metallic surface under investigation. The corrosion potential can be derived from the Volta potential difference using a suitable calibration technique. One of the main advantages of the SKP technique is its nondestructive nature, while one disadvantage could be the fact that only potentials can be monitored. The SKP has been mainly applied to the study of polymer-coated metal surfaces, because it allows the detection of distinct potential differences between delaminated and intact areas (Ref 73). The SKP technique has been used to monitor the progress of filiform corrosion on a painted aluminum panel (Ref 65). The SKP has been used to map the potential distribution under small drops of electrolyte in the classic Evans drop experiment (Ref 74, 75). Small drops of NaCl were placed on carbon steel, galvanized steel, stainless steel, and an aluminum alloy. In the absence of chromate, local anodes and cathodes were established in a short time. When dichromate was added to NaCl, very uniform potential distribution was found for stainless steel and Al 2024. The main driving force for the development of local anodes was considered to arise from oxygen reduction in very thin films of electrolyte surrounding the NaCl drop (Ref 74). The SKP technique has also been applied to the study of atmospheric corrosion processes (Ref 76).

References cited in this section 8. R. Baboian, Ed., Electrochemical Techniques for Corrosion Engineers, NACE, 1986 65. J.H.W. de Wit, D.H. van der Weijde, A.J. de Jong, F. Blekkenhorst, and S.D. Meijers, Mat. Sci. Forum, Vol 289–292, 1998, p 69 66. K.R. Trethewey, D.A. Sargeant, D.J. Marsh, and A.A. Tamimi, Corros. Sci., Vol 35, 1993, p 127

67. F.J. Maile, T. Schauer, and C.D. Eisenbach, Prog. Org. Coatings, Vol 38, 2000, p 117 68. F.J. Maile, T. Schauer, and C.D. Eisenbach, Farbe + Lack, Vol 104 (No. 11), 1998, p 58 69. F. Zou, C. Barreau, R. Hellouin, D. Quantin, and D. Thierry, Mat. Sci. Forum, Vol 289–292, 1998, p 83 70. G. Chen, R.J. Palmer, and D.C. White, Biodegrad., Vol 8, 1997, p 189 71. M. Stratmann, Corros. Sci., Vol 27, 1987, p 869 72. M. Stratmann, Corros. Sci., Vol 32, 1991, p 467 73. M. Stratmann, H. Streckel, and R. Feser, Farbe + Lack, Vol 97, 1991, p 9 74. C. Chen, C.B. Breslin, and F. Mansfeld, Mater. Corros., Vol 49, 1998, p 1 75. U.R. Evans, Met. Ind., Vol 29, 1926, p 481 76. M. Stratmann and H. Streckel, Corros. Sci., Vol 30, 1990, p 681, 697, 715


Miscellaneous Techniques A number of electrochemical techniques have been developed that allow evaluation of processes such as hydrogen embrittlement, especially of steels, or the susceptibility of metal and alloys to localized corrosion, such as pitting. Certain standard laboratory tests, such as the anodized aluminum corrosion test, the impedance test for anodized aluminum, the electrolytic corrosion test, as well as the paint adhesion on a scribed surface (PASS) test, are used mainly as quality-control tests. Hydrogen Permeation Technique. The Devanathan-Stachurski method can be used to study parameters associated with the hydrogen embrittlement tendencies of metals and alloys (Ref 77). This method is described in ASTM G 148, “Standard Practice for Evaluation of Hydrogen Uptake, Permeation, and Transport in Metals by an Electrochemical Technique.” The technique involves placing a thin metal foil between two electrochemical cells—a hydrogen charging cell and a hydrogen oxidation cell (Fig. 31). Hydrogen is generated on the side of the metal foil facing the charging cell by application of a sufficiently negative potential. For characterization of the hydrogen transport in the bulk of the material under investigation, galvanostatic charging is preferred. The hydrogen that diffuses through the foil is oxidized on the side of the foil that faces the oxidation cell. The potential of the foil is controlled at a sufficiently positive potential to ensure that the kinetics of the oxidation of hydrogen atoms are limited by the flux of hydrogen atoms. The total oxidation current is monitored as a function of time. It is the sum of the hydrogen oxidation current, which is the permeation current, and the background current due to dissolution of the metal foil. The latter needs to be kept small by choosing a proper electrolyte in the oxidation cell. For ferrous materials, an alkaline solution, such as 0.1 M NaOH, is used to ensure passive behavior.

Fig. 31 Electrochemical hydrogen permeation cell assembly and measuring apparatus. (a) Schematic. (b) Polytetrafluoroethylene (PTFE) hydrogen permeation cell. Source: ASTM G 148 For a known foil thickness, L, the steady-state permeation current, Jss, gives information concerning the subsurface concentration, CH, at the charging surface: Jss = CHDl/L

(Eq 13)

where Dl is the lattice diffusion coefficient of hydrogen. ASTM G 148 gives a detailed description of the procedure for the analysis of the results of electrochemical hydrogen permeation results obtained from steadystate permeation current measurements or from permeation transient measurements. The barnacle electrode assembly allows determination of the amount of mobile hydrogen atoms in a metallic structure that had been exposed previously to an aqueous or gaseous environment. In this case, only the hydrogen oxidation cell is used (Ref 78, 79). Standard Laboratory Tests. Numerous ac and dc electrochemical methods have been used to study the performance and quality of protective coatings, including passive films on metallic substrates, and to evaluate the effectiveness of various surface pretreatments. Several of these methods are discussed subsequently. The Anodized Aluminum Corrosion Test. The Ford anodized aluminum corrosion test (FACT) (Ref 80) involves cathodic polarization of the anodized aluminum surface by using a small, cylindrical, glass clamp-on cell and a special 5% NaCl solution containing cupric chloride (CuCl2) acidified with acetic acid. A large voltage is applied across the cell by using a platinum auxiliary electrode. The alkaline conditions created by the cathodic polarization promote dissolution at small defects in the anodized aluminum surface. As the coating resistance is decreased, more current begins to flow, and the voltage decreases. The cell voltage (auxiliary electrode to test specimen voltage) is monitored for 3 min, and the parameter “cell voltage multiplied by time” is recorded. ASTM B 538, “Standard Method of FACT (Ford Anodized Aluminum Corrosion Test) Testing,” describes the method in greater detail. A similar test, known as the cathodic breakdown test, involves cathodic polarization to -1.6 versus SCE for a period of 3 min in acidified NaCl. The test was designed for anodized aluminum alloys, because the alkaline conditions created by the large applied currents promote the formation of corroding spots at defects in the anodized film. The Electrolytic Corrosion Test. This test was designed for electrodeposits of, principally, nickel and chromium on less noble metals, such as zinc or steel (Ref 81). Special solutions are used, and the metal is polarized to +0.3 V versus SCE. The metal is taken through cycles of 1 min anodic polarization and 2 min unpolarized condition. An indicator solution is then used to detect the presence of pits that penetrate to the substrate. Each exposure cycle is assumed to simulate one year of exposure under atmospheric corrosion conditions. ASTM B 627, “Standard Method for Electrolytic Corrosion Testing (EC Test),” describes the method in greater detail.

The PASS test involves cathodic polarization of a small portion of a painted metal. The area exposed contains a scribed line that exposes a line of underlying bare metal. The sample is cathodically polarized for 15 min in 5% NaCl. At the end of this period, the amount of delaminated coating is determined using an adhesive-tape pulling procedure. The Impedance Test for Anodized Aluminum. ASTM B 457, “Standard Method for Measurement of the Impedance of Anodic Coatings on Aluminum,” is used to study the efficiency of the sealing process for anodized aluminum alloys (Ref 82, 83). In this sense, the test is similar to the FACT test, except that this method uses a 1 Vrms, 1 kHz signal source from an impedance bridge to determine the impedance of sealed, anodized aluminum. The impedance of the sealed, anodized aluminum alloy is determined in kΩ. The test area is defined with a portable cell, and a platinum or stainless steel auxiliary electrode is typically used. The sample is immersed in 3.5% NaCl. In contrast to the methods discussed previously, this quality-control test is essentially nondestructive and does not accelerate the corrosion process. The Critical Pitting Temperature. The susceptibility of a material to pitting is usually evaluated based on the pitting potential, Epit, that can be determined using a potentiostatic or potentiodynamic (pitting scan) technique. An alternative approach is based on the critical pitting temperature (CPT) that is described in ASTM G 150, “Standard Test Method for Electrochemical Critical Temperature Testing of Stainless Steels.” The CPT is defined as the lowest temperature at which stable pit propagation occurs. The test method determines the potential-independent critical pitting temperature using a potentiostatic technique with a temperature scan. A specific sample holder is recommended that eliminates the occurrence of crevice corrosion. The test sample is exposed to a 1M NaCl solution that is initially maintained at 0 °C. The solution is then heated at a rate of 1 °C/min. The sample is polarized to a potential above the Epit range during the entire temperature scan. The current is monitored during the temperature scan, and the CPT is defined as the temperature at which the current density exceeds 100 μA/ cm2. The occurrence of pitting is confirmed visually after the test. The standard parameters recommended in ASTM G 150 are suitable for determining the CPT of austenitic stainless steels and other related alloys.

References cited in this section 77. M.A.V. Devanathan and Z. Stachurski, Proc. R. Soc. (London) A, Vol 270, 1962, p 90 78. D.A. Berman, J.J. DeLuccia, and F. Mansfeld, Met. Prog., Vol 115, 1979, p 58 79. F. Mansfeld, S. Jeanjaquet, and D.K. Roe, Mater. Perform., Vol 21 (No. 2), 1982, p 36 80. J. Stone, H.A. Tuttle, and H.N. Bogart, Plating, Vol 43, 1965, p 877 81. R.L. Saur and R.P. Basco, Plating, Vol 53, 1966, p 33, 320, 981 82. E.T. Englehart and G. Sowinski, Jr., SAE J., Vol 72, 1974, p 51 83. E.T. Englehart and D.G. George, Mater. Prot. (China), Vol 3, 1964, p 25 F. Mansfeld, Electrochemical Methods of Corrosion Testing, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 446–462 Electrochemical Methods of Corrosion Testing Florian Mansfeld, University of Southern California

References 1. C. Wagner and W. Traud, Z. Elektrochemie, Vol 44, 1938, p 391

2. H. Kaesche, The Corrosion of Metals, NACE, 1985 3. D.A. Jones, Principles and Prevention of Corrosion, MacMillan Publishing Co., 1992 4. F. Mansfeld, The Polarization Resistance Technique for Measuring Corrosion Currents, Advancements in Corrosion Science and Technology, Vol 6, M.G. Fontana and R.W. Staehle, Ed., Plenum Press, 1976, p 163 5. M. Stern and A.L. Geary, J. Electrochem. Soc., Vol 104, 1957, p 56 6. R. Baboian, Ed., Electrochemical Techniques for Corrosion, NACE, 1977 7. Electrochemical Corrosion Testing, STP 727, U. Bertocci and F. Mansfeld, Ed., ASTM, 1979 8. R. Baboian, Ed., Electrochemical Techniques for Corrosion Engineers, NACE, 1986 9. R. Baboian, Ed., Corrosion Tests and Standards—Application and Interpretation, ASTM, 1995 10. R.W. Revie, Ed., Uhlig's Corrosion Handbook, 2nd ed., Electrochemical Society Series, WileyInterscience, 2000 11. B.J. Little, P.A. Wagner, and F. Mansfeld, Corrosion Testing Made Easy—Microbiologically Influenced Corrosion, B.C. Syrett, Ed., NACE, 1997 12. N.G. Thompson and J.H. Payer, Corrosion Testing Made Easy—dc Electrochemical Methods, B.C. Syrett, Ed., NACE, 1998 13. R. Cottis and S. Turgoose, Corrosion Testing Made Easy—Electrochemical Impedance and Noise, B.C. Syrett, Ed., NACE, 1999 14. F. Mansfeld, Corrosion, Vol 44, 1988, p 856 15. G.G. Geesey, I. Beech, P.J. Bremer, B.J. Webster, and D.B. Wells, Biocorrosion, Biofilms II: Process Analysis and Applications, J.D. Bryers, Ed., Wiley, 2000, p 281 16. F. Mansfeld and M.W. Kendig, Corrosion, Vol 41, 1985, p 490 17. F. Mansfeld, Corrosion, Vol 32, 1976, p 143 18. F. Mansfeld, M.W. Kendig, and S. Tsai, Corros. Sci., Vol 22, 1982, p 455 19. F. Mansfeld and B. Little, Corros. Sci., Vol 32, 1991, p 247 20. A. Naguib and F. Mansfeld, Electrochim. Acta, Vol 37, 2002, p 2319 21. G. Lauer and F. Mansfeld, Corrosion, Vol 26, 1970, p 504 22. F. Mansfeld and R.V. Inman, Corrosion, Vol 35, 1979, p 21 23. F. Mansfeld, Corrosion, Vol 29, 1973, p 403 24. F. Mansfeld, Corrosion, Vol 27, 1971, p 436 25. F. Mansfeld, D.H. Hengstenberg, and J.V. Kenkel, Corrosion, Vol 30, 1974, p 343

26. F. Mansfeld and J.V. Kenkel, Corros. Sci., Vol 15, 1975, p 183, 239 27. D.A. Jones, Corrosion, Vol 40, 1984, p 181 28. F. Mansfeld, L.T. Han, and C.C. Lee, J. Electrochem. Soc., Vol 143, 1996, p L286 29. F. Mansfeld, L.T. Han, C.C. Lee, C. Chen, G. Zhang, and H. Xiao, Corros. Sci., Vol 39, 1997, p 255 30. F. Mansfeld, L.T. Han, C.C. Lee, and G. Zhang, Electrochim. Acta, Vol 43, 1998, p 2933 31. F. Mansfeld, H. Xiao, L.T. Han, and C.C. Lee, Prog. Org. Coatings, Vol 30, 1997, p 89 32. D.A. Eden, “Electrochemical Noise—The First Two Octaves,” Paper 386, Corrosion/ 98, NACE, 1998 33. U. Bertocci, C. Gabrielli, F. Huet, and M. Keddam, J. Electrochem. Soc., Vol 144, 1997, p 31 34. U. Bertocci, C. Gabrielli, F. Huet, M. Keddam, and P. Rousseau, J. Electrochem. Soc., Vol 144, 1997, p 37 35. F. Mansfeld, Z. Sun, E. Speckert, and C.H. Hsu, “Electrochemical Noise Analysis (ENA) for Active and Passive Systems,” Paper 00418, Corrosion/2000, NACE, 2000 36. F. Mansfeld, Z. Sun, C.H. Hsu, and A. Nagiub, Corros. Sci., Vol 43, 2001, p 341 37. F. Mansfeld and Z. Sun, Corrosion, Vol 55, 1999, p 915 38. C.C. Lee and F. Mansfeld, Corros. Sci., Vol 40, 1998, p 959 39. C. Gabrielli, F. Huet, and M. Keddam, Electrochim. Acta, Vol 31, 1986, p 1025 40. J.R. Scully, Corrosion, Vol 56, 2000, p 199 41. F. Mansfeld, H. Shih, H. Greene, and C.H. Tsai, Analysis of EIS Data for Common Corrosion Processes, STP 1188, ASTM, 1993, p 37 42. F. Mansfeld, J. Appl. Electrochem., Vol 25, 1995, p 187 43. F. Mansfeld and M. Kendig, Electrochemical Impedance Tests for Protective Coatings, STP 866, ASTM, 1985, p 122 44. F. Mansfeld, M. Kendig, and S.Tsai, Corrosion, Vol 38, 1982, p 478 45. M. Kendig, F. Mansfeld, and S. Tsai, Corros. Sci., Vol 23, 1983, p 317 46. J. Titz, G.H. Wagner, H. Spaehn, M. Ebert, K. Juettner, and W.J. Lorenz, Corrosion, Vol 46, 1990, p 221 47. M. Kendig and F. Mansfeld, Mat. Res. Soc. Symp. Proc., Vol 115, 1988, p 193 48. F. Mansfeld, C.C. Lee, L.T. Han, B.J. Little, P. Wagner, R. Ray, and J. Jones-Meehan, “The Impact of MIC on Protective Polymer Coatings,” Final Report, Contract N00014- 94-1-0026, The Office of Naval Research, Aug 1998 49. J. Hitzig, K. Juettner, W.J. Lorenz, and W. Paatsch, Corros. Sci., Vol 24, 1984, p 945

50. J. Hitzig, K. Juettner, W.J. Lorenz, and W. Paatsch, J. Electrochem. Soc., Vol 133, 1986, p 133 51. F. Mansfeld and M.W. Kendig, J. Electrochem. Soc., Vol 135, 1988, p 828 52. F. Mansfeld, S. Lin, S. Kim, and H. Shih, Corros. Sci., Vol 27, 1987, p 997 53. F. Mansfeld and H.Shih, J. Electrochem. Soc., Vol 135, 1987, p 332 54. F. Mansfeld, S. Lin, S. Kim, and H. Shih, Electrochim. Acta, Vol 34, 1989, p 1123 55. F. Mansfeld, S. Lin, S. Kim, and H. Shih, J. Electrochem. Soc., Vol 137, 1990, p 78 56. F. Mansfeld, H. Shih, H. Greene, and C.H. Tsai, Detection and Monitoring of Localized Corrosion with EIS, STP 1188, ASTM, 1993, p 297 57. F. Mansfeld and M. Kendig, Corrosion, Vol 37, 1981, p 545 58. M. Dubiel, C.H. Hsu, C.C. Chien, F. Mansfeld, and D.K. Newman, Microbial Iron Respiration Can Protect Steel from Corrosion, Appl. Environ. Microbiol., submitted for publication 59. V. Cihal, A. Desestret, M. Froment, and G.H. Wagner, Proc. Conf. European Federation of Corrosion (Paris), European Federation of Corrosion, 1973, p 249 60. K. Osozawa, K. Bohnenkamp, and H.J. Engell, Corros. Sci., Vol 6, 1966, p 421 61. W.L. Clarke, R.L. Cowan, and W.L. Walker, in Intergranular Corrosion of Stainless Alloys, STP 656, ASTM, 1978, p 99 62. M. Akashi, T. Kawamoto, and F. Umemura, Corros. Eng. (Jpn.), Vol 29, 1980, p 163 63. A.P. Majidi and M.A. Streicher, Corrosion, Vol 42, 1984, p 584 64. H. Shih and F. Mansfeld, STP 1154, ASTM, 1992, p 174 65. J.H.W. de Wit, D.H. van der Weijde, A.J. de Jong, F. Blekkenhorst, and S.D. Meijers, Mat. Sci. Forum, Vol 289–292, 1998, p 69 66. K.R. Trethewey, D.A. Sargeant, D.J. Marsh, and A.A. Tamimi, Corros. Sci., Vol 35, 1993, p 127 67. F.J. Maile, T. Schauer, and C.D. Eisenbach, Prog. Org. Coatings, Vol 38, 2000, p 117 68. F.J. Maile, T. Schauer, and C.D. Eisenbach, Farbe + Lack, Vol 104 (No. 11), 1998, p 58 69. F. Zou, C. Barreau, R. Hellouin, D. Quantin, and D. Thierry, Mat. Sci. Forum, Vol 289–292, 1998, p 83 70. G. Chen, R.J. Palmer, and D.C. White, Biodegrad., Vol 8, 1997, p 189 71. M. Stratmann, Corros. Sci., Vol 27, 1987, p 869 72. M. Stratmann, Corros. Sci., Vol 32, 1991, p 467 73. M. Stratmann, H. Streckel, and R. Feser, Farbe + Lack, Vol 97, 1991, p 9 74. C. Chen, C.B. Breslin, and F. Mansfeld, Mater. Corros., Vol 49, 1998, p 1

75. U.R. Evans, Met. Ind., Vol 29, 1926, p 481 76. M. Stratmann and H. Streckel, Corros. Sci., Vol 30, 1990, p 681, 697, 715 77. M.A.V. Devanathan and Z. Stachurski, Proc. R. Soc. (London) A, Vol 270, 1962, p 90 78. D.A. Berman, J.J. DeLuccia, and F. Mansfeld, Met. Prog., Vol 115, 1979, p 58 79. F. Mansfeld, S. Jeanjaquet, and D.K. Roe, Mater. Perform., Vol 21 (No. 2), 1982, p 36 80. J. Stone, H.A. Tuttle, and H.N. Bogart, Plating, Vol 43, 1965, p 877 81. R.L. Saur and R.P. Basco, Plating, Vol 53, 1966, p 33, 320, 981 82. E.T. Englehart and G. Sowinski, Jr., SAE J., Vol 72, 1974, p 51 83. E.T. Englehart and D.G. George, Mater. Prot. (China), Vol 3, 1964, p 25

T. Yau, Immersion Testing, Corrosion: Fundamentals, Testing, and Protection,, Vol 13A, ASM Handbook, ASM International, 2003, p 463–469

Immersion Testing Te-Lin Yau, Te-Lin Yau Consultancy

Introduction IMMERSION TESTING is the most frequently conducted test for evaluating the corrosion of metals in aqueous solutions. It appears simple: Totally immerse a test specimen in a corrosive solution for a period of time and then remove the specimen for examination. It is sometimes referred to as a “quick and dirty” test. However, a number of factors must be considered to achieve specific goals and to ensure adequate reproducibility of test results. The actual conditions of the test are usually determined by the nature of the problem at hand, the ingenuity of the investigator, and the budget of the test program. The following information is presented as a guide, so that some of the pitfalls of such testing may be avoided. The methods and procedures described include basic factors that must be considered in sophisticated test procedures as well as in simple tests. Primary consensus standards for immersion corrosion testing of metals have been developed by ASTM International and NACE International. Applicable ASTM standards are given in Ref 1 to 9. The NACE method is given in Ref 10. For proper planning of the test and interpretation of the test results, the specific influences of the following variables must be considered: solution composition, temperature, aeration, volume, velocity, and waterline effects; specimen surface preparation; method of immersion of specimens; duration of test; and method of cleaning specimens at the conclusion of the exposure. In most cases, immersion tests are conducted to determine the corrosion rates of metals in a given environment. However, by employing specifically designed specimens and/or environments, immersion tests can also be conducted to evaluate the resistance of the metal to pitting, crevice corrosion, galvanic corrosion, hydrogen embrittlement, erosion, and stress-corrosion cracking. Sections of this Volume devoted to these topics have details on the specific testing methods.

References cited in this section 1. “Practice for Detecting Susceptibility to Intergranular Attack in Austenitic Stainless Steels,” A 262, Annual Book of ASTM Standards, ASTM 9. “Standard Test Method for Visual Assessment of Exfoliation Corrosion Suscepti[chbility of 5xxx Series Aluminum Alloys,”[fj G 66, Annual Book of ASTM Standards, ASTM 10. “Laboratory Corrosion Testing of Metals,” NACE TM-0169-2000, National Association of Corrosion Engineers

T. Yau, Immersion Testing, Corrosion: Fundamentals, Testing, and Protection,, Vol 13A, ASM Handbook, ASM International, 2003, p 463–469 Immersion Testing Te-Lin Yau, Te-Lin Yau Consultancy

Total Immersion Composition of Solution. Test solutions can be: (a) accurately prepared by combining chemicals according to the specifications of the Committee on Analytical Reagents of the American Chemical Society with distilled water, (b) naturally occurring solutions, or (c) samples of plant process solutions. The composition of the test solutions should be controlled to the fullest extent possible. They should be described as completely and accurately as possible in the test report. It helps to include information such as color, pH, and specific gravity before and after the test. Evaporation losses can be controlled by a constant level device or by frequent addition of appropriate solution to maintain the original volume within ±1%. The use of a reflux condenser ordinarily, but not always, precludes the necessity of adding to the original kettle charge. Certain chemicals are too volatile to be quickly condensed. The composition of the test solution sometimes changes as a result of catalytic decomposition or by reaction with the test specimens. For example, hydrogen peroxide may catalytically decompose, and hydrofluoric acid is quickly consumed by reacting with reactive metals such as titanium and zirconium. These changes should be determined if possible. Where required, the exhausted constituents should be added or a fresh solution provided during the course of the test. If problems are suspected, the composition of the test solutions should be checked by analysis at the end of the test to determine the extent of change in composition that took place from evaporation or reactions. Corrosion products from the specimen may influence the corrosion rate of the metal or the corrosion rate of different metals exposed at the same time. For example, the accumulation of cupric ions in the testing of copper alloys in intermediate strength of sulfuric acid accelerates the corrosion of copper alloys, as compared to the rates that are found if the corrosion products are continually removed. Cupric ions may also exhibit a passivating effect on stainless steel specimens exposed at the same time. It is generally advisable to expose only the same type of alloys in the testing apparatus. Alloys in the same group may be quite different in chemical compositions. For example, there are copper-containing stainless steels, such as some of the precipitationhardening and duplex stainless steels, and there are copper-free stainless steels. Temperature of Solution. Temperature is a critical factor in corrosion. Accordingly, the temperature of the test solution and, if different, the temperature of the test specimen must be known. In many cases, corrosion increases rapidly with temperature. The corrosion of type 303 stainless steel by 65% nitric acid is shown in Table 1 (Ref 10). Corrosion increases from a low rate of less than 0.5 mm/yr (20 mils/yr) to over 100 mm/yr (4000 mils/yr) as a result of increasing the temperature 65 °C (120 °F). For some corrosion systems, the thermal effect is more complex. For example, the corrosion rate of Monel alloy in air- saturated sulfuric acid solutions increases sharply with rising temperature to a maximum at 80 °C (175 °F) and then decreases with a further increase in temperature. This is because of the decrease in dissolved oxygen in sulfuric acid, as indicated in Fig. 1 (Ref 11, 12). Table 1 Effect of temperature on corrosion of type 303 stainless steel by 65% HNO3 Temperature °C °F Up to 120 250 260 125 280 135 320 160 330 165

Corrosion rate mm/yr mils/yr A2 Increased Increased A1 < B < A2 Increased

Comparison for corrosion damage A1 on fresh metal in the time interval from 0 to 1 with corrosion damage B on fresh metal for the unit time interval from t to t + 1 shows the magnitude and direction of change in the corrosivity of the liquid that may have occurred during the total time of the test. Correspondingly, comparison of A2 with B, where A2 is the corrosion damage calculated by subtracting At from At+1, shows the magnitude and direction of change in corrodibility of the metal specimen during the test. These comparisons can be taken as criteria for the changes and are tabulated in Table 2. Also given in Table 2 are the criteria for all possible combinations of changes in the corrosivity of the liquid and corrodibility of the metal. The additional information obtained on occurrences in the course of the test justifies the additional effort involved. Table 3 lists the data obtained from a planned interval test of a carbon steel specimen in an aluminum chloride (AlCl3)/antimony trichloride (SbCl3) solution.

Table 3 Planned interval corrosion test Duplicate strips of low-carbon steel (19 × 75 mm, or × 3 in.) were immersed in 200 mL of 10% AlCl3-90% SbCl3 mixture through which dried hydrogen chloride gas was slowly bubbled at atmospheric pressure. Temperature: 90 °C (195 °F) Penetration Apparent corrosion rate Value Interval, days Weight loss, mg μm mils mm/yr mils/yr 0–1 1080 42.9 1.69 15.7 620 A1 0–3 1430 56.9 2.24 6.8 270 At 0–4 1460 58.2 2.29 5.3 210 At+1 3–4 70 2.8 0.11 1.02 40 B Calculated 3–4 30 1.3 0.05 0.46 18 A2 Analysis: A2 < B < A1, or 0.05 < 0.11 < 1.69. Conclusions: Liquid markedly decreased in corrosiveness during the test, and the formation of a partially protective scale on the steel was indicated; that is, metal became less corrodible. The causes for the changes in corrosion rate as a function of time are not given by the planned interval test criteria. The corrosivity of the liquid may decrease as a result of corrosion during the course of a test because of the reduction in concentration of the corrosive agent, the depletion of a corrosive contaminant, the formation of inhibiting products, or other metal-catalyzed changes in the liquid. The corrosivity of the liquid may increase because of the formation of autocatalytic products, the destruction of corrosion-inhibiting substances, or other catalyzed changes in the liquid. Changes in the corrosiveness of the liquid may also arise from changes in composition that would occur under the test conditions, even in the absence of metal. To determine if the latter effect occurs, an identical test is run without test strips for the total time, t. Test strips are then added, and the test is continued for the unit time interval. Comparison with A1 of corrosion damage from this test shows if the corrosive character of the liquid changes significantly in the absence of metal. The corrodibility of the metal in a test may decrease as a function of time because of the formation of protective scale or the removal of a less resistant surface layer of metal. Metal corrodibility may increase because of the formation of corrosion-accelerating scale or the removal of a more resistant surface layer of metal. Indications of the causes of changes in corrosion rate can often be obtained from close observation of tests and corroded specimens as well as from special supplementary tests designed to reveal effects that may be involved. Changes in liquid corrosiveness are not a factor in most plant tests that consist of once- through runs or where large ratios of solution volume to specimen area are involved. If the effect of corrosion on the mechanical properties of the metal or alloy is under consideration, a set of unexposed specimens is needed for comparison purposes. Lengthy corrosion tests are generally not necessary to obtain accurate corrosion rates from materials that undergo severe corrosion. However, there are cases in which this assumption is not valid. For example, lead exposed to sulfuric acid initially corrodes at an extremely high rate while building a protective film; the rates then decrease considerably, so that further corrosion is negligible. The phenomenon of the formation of a protective film is observed with many corrosion-resistant materials. Therefore, short tests on such materials would indicate a high corrosion rate and would be completely misleading. Short-term tests can also give misleading results on alloys that form passive films, such as stainless steels. With borderline conditions, a prolonged test may be needed to permit the breakdown of the passive film and the subsequent more rapid attack. Consequently, tests conducted for long periods are considerably more realistic than those conducted for short durations. This statement must be qualified by stating that corrosion should not proceed to the point at which the original specimen size or the exposed area is drastically reduced or the metal is perforated. If anticipated corrosion rates are moderate or low, the following equation gives the suggested tests duration:

For example, where the corrosion rate is 10 mils/yr (0.25 mm/yr), the test should run for at least 200 h. This method of estimating test duration is useful as an aid in deciding, after a test has been made, whether or not it is desirable to repeat the test for a longer period.

References cited in this section 4. “Practice for Preparing, Cleaning, and Evaluating Corrosion Test Specimens,” G 1, Annual Book of ASTM Standards, ASTM 7. “Practice for Laboratory Immersion Corrosion Testing of Metals,” G 31, Annual Book of ASTM Standards, ASTM 10. “Laboratory Corrosion Testing of Metals,” NACE TM-0169-2000, National Association of Corrosion Engineers 14. M.G. Fontana, “Corrosion Testing,” Lesson 8 in home study and extension courses, American Society for Metals, 1969 22. T.-L. Yau, J.A. Andrews, R. Henson, and D.R. Holmes, Practice for Conducting Corrosion Coupon Tests on Zirconium and Its Alloys, Corrosion Tests and Evaluation: Silver Anniversary Volume, STP 1000, R. Baboian and S.W. Dean, Ed., ASTM, 1990, p 303–311


Cleaning Corrosion Test Specimens After corrosion testing, specimens should be properly cleaned to remove bulky deposits and corrosion products but not the metals. This must be done as soon as possible, so that the measured corrosion time interval is accurate. Corrosion may be continuing under the damp corrosion products until they are removed. Various mechanical, electrolytic, and chemical cleaning methods have been developed to clean different metals and alloys (Ref 4).

Reference cited in this section 4. “Practice for Preparing, Cleaning, and Evaluating Corrosion Test Specimens,” G 1, Annual Book of ASTM Standards, ASTM


Calculation of Corrosion Rate

The average corrosion rate may be obtained as follows: Corrosion rate = (K × W)/(A × T × D) where K is a constant, T is the time of exposure in hours to the nearest 0.01 h, A is the area in cm2 to the nearest 0.01 cm2, W is the mass loss in grams to the nearest 1 mg, and D is the density in g/cm3. Several units are used to express corrosion rates. Using the previously mentioned units for T, A, W, and D, the corrosion rate can be calculated in a variety of units, with the following appropriate value of K: Corrosion rate units desired Constant, K 3.45 × 106 Mils per year (mils/yr) 3.45 × 103 Inches per year (in./yr) 8.76 × 104 Millimeters per year (mm/yr) 8.76 × 107 Micrometers per year (μm/yr) Milligrams per square decimeter per day (mdd) 2.40 × 106 × D Adapted from Ref 4 When expressing corrosion rates in years, the corrosionist must keep in mind the underlying assumption that the corrosion rate is uniform. When based on test data measured in hours, the extrapolation to years may not be valid. Also, by measuring the weight loss, an average loss of thickness is calculated. This may not be the most relevant characteristic.

Reference cited in this section 4. “Practice for Preparing, Cleaning, and Evaluating Corrosion Test Specimens,” G 1, Annual Book of ASTM Standards, ASTM


Reporting the Data The importance of reporting all data as completely as possible cannot be overemphasized. Expansion of the testing program in the future or correlating the results with tests of other investigators is possible only if all pertinent information is properly recorded, so the veracity of the data can be assured. The following checklist is a recommended guide for items that should be included in a standard report of immersion tests: • • • • • • • • • • • •

Name of personnel and laboratory involved Date Identification of specimen material and number Form and metallurgical conditions of specimens (composition, quality grade) Cite standard methods employed (industry, government, or laboratory standards and guides) Treatment used to prepare specimens Visual-appearance remarks of each specimen before and after the corrosion testing Weights and dimensions of specimens before and after the corrosion testing Corrosive media and concentration (and any changes during test) Volume of test solution The color, pH, and specific gravity of the test solution before and after the corrosion testing Temperature (maximum, minimum, and average)

• • • • • • • • •

Aeration (describe conditions or technique) Agitation (describe conditions or technique) Type of apparatus used for test. Include identification of measuring devices for auditing calibrations in accordance with established laboratory procedure. Duration of each test Number of specimens of each material tested, and whether specimens were tested separately or which specimens were tested in the same container Method used to clean specimens after exposure, and the extent of any error expected by this treatment. Use of a documented standard method is desirable. Evaluation of attack if other than general, such as crevice corrosion under support rod, pit depth and distribution, and results of microscopic examination or bend tests Corrosion rates for each specimen Minor occurrences or deviations from the proposed test program often can have significant effects and should be reported, if known.


Acknowledgment Whereas this article was substantially revised, acknowledgment is made that it was adapted in part from the original article, “Immersion Testing,” by D.O. Sprowls, Corrosion, Vol 13, ASM International, 1987, p 220.


References 1. “Practice for Detecting Susceptibility to Intergranular Attack in Austenitic Stainless Steels,” A 262, Annual Book of ASTM Standards, ASTM 2. “Standard Guide for Descaling and Cleaning Titanium and Titanium Alloy Surfaces,” B 600, Annual Book of ASTM Standards, ASTM 3. “Standard Guide for Descaling and Cleaning Zirconium and Zirconium Alloy Surfaces,” B 614, Annual Book of ASTM Standards, ASTM 4. “Practice for Preparing, Cleaning, and Evaluating Corrosion Test Specimens,” G 1, Annual Book of ASTM Standards, ASTM

5. “Guide for Conducting Corrosion Tests in Field Applications,” G4, Annual Book of ASTM Standards, ASTM 6. “Guide for Applying Statistics to Analysis of Corrosion Data,” G 16, Annual Book of ASTM Standards, ASTM 7. “Practice for Laboratory Immersion Corrosion Testing of Metals,” G 31, Annual Book of ASTM Standards, ASTM 8. “Guide for Examination and Evaluation of Pitting Corrosion,” G 46, Annual Book of ASTM Standards, ASTM 9. “Standard Test Method for Visual Assessment of Exfoliation Corrosion Suscepti[chbility of 5xxx Series Aluminum Alloys,”[fj G 66, Annual Book of ASTM Standards, ASTM 10. “Laboratory Corrosion Testing of Metals,” NACE TM-0169-2000, National Association of Corrosion Engineers 11. F.A. Champion, Corrosion Testing Procedures, 2nd ed., John Wiley & Sons, 1965 12. O.B.J. Fraser, D.E. Ackerman, and J.W. Sands, Ind. Eng. Chem., Vol 19, 1927, p 332 13. P.E. Francis and A.D. Mercer, Corrosion of a Mild Steel in Distilled Water and Chloride Solutions: Development of a Test Method, Laboratory Corrosion Tests and Standards, STP 866, G.S. Haynes and R. Baboian, Ed., ASTM, 1985, p 184–196 14. M.G. Fontana, “Corrosion Testing,” Lesson 8 in home study and extension courses, American Society for Metals, 1969 15. W.H. Ailor, Ed., Handbook on Corrosion Testing and Evaluation, John Wiley & Sons, 1971 16. T.-L. Yau, Corrosion Test Loop, Laboratory Corrosion Tests and Standards, STP 866, G.S. Haynes and R. Baboian, Ed., ASTM, 1985, p 36–47 17. G.D. Bengough, U.R. Evand, T.P. Hoar, and F. Wormweff, Chem. Ind., Nov 1938 18. “Standard Test Method for Corrosion Testing of Products of Zirconium, Hafnium, and Their Alloys in Water at 680 °F or in Steam at 750 °F,” G 2, Annual Book of ASTM Standards, ASTM 19. “Standard Recommendation Practice for Alternate Immersion Stress Corrosion Testing in 3.5% Sodium Chloride Solution,” G 44, Annual Book of ASTM Standards, ASTM 20. “Standard Specification for Reagent Water,” D 1193, Annual Book of ASTM Standards, ASTM 21. “Standard Specification for Substitute Ocean Water,” D 1141, Annual Book of ASTM Standards, ASTM 22. T.-L. Yau, J.A. Andrews, R. Henson, and D.R. Holmes, Practice for Conducting Corrosion Coupon Tests on Zirconium and Its Alloys, Corrosion Tests and Evaluation: Silver Anniversary Volume, STP 1000, R. Baboian and S.W. Dean, Ed., ASTM, 1990, p 303–311


Selected References • • • •

R. Baboian, Ed., Manual 20: Corrosion Tests and Standards: Application and Interpretation, ASTM, 1995 R. Baboian and S.W. Dean, Ed., Corrosion Testing and Evaluation: Silver Anniversary Volume, STP 1000, ASTM, 1990 G. Cragnolino and N. Sridhar, Ed., Application of Accelerated Corrosion Tests to Service Life Prediction of Materials, STP 1194, ASTM, 1995 E.D. Verink, Corrosion Testing Made Easy: The Basics, National Association of Corrosion Engineers, 1993

R. Singleton, Cabinet Testing, Corrosion: Fundamentals, Testing, and Protection ,Vol 13A, ASM Handbook, ASM International, 2003, p 470–477

Cabinet Testing Raymund Singleton, The Singleton Corporation

Introduction SALT SPRAY (FOG) TESTS have been used for over 100 years as accelerated tests for determining both the corrodibility of nonferrous and ferrous metals and the degree of protection afforded by both inorganic and organic coatings on metallic base (Ref 1). The tests are utilized most often to compare the relative performance of metals and coatings to a known and retained standard for quality control purposes. The tests can be an indicator of the consistency of a production process and/or materials utilized. There has been extensive discussion of the salt fog test since its inception because of the reproducibility variances and the questionable correlation of results as related to actual “inservice” performance. This can be a result of the misapplication of the salt fog test with the expectation that the test atmosphere simulates a particular environment or service condition. Salt spray tests were not intended for this purpose, except in rare instances where corroborating data has been obtained to support such application. The test actually represents the application of a standardized amount of corrosive activity for comparing the relative performance of materials and/or coatings. Many revisions and improvements to the salt spray test procedures, salt spray test chambers, and apparatus have been made over the years. A significant amount of the contributions result from efforts of ASTM International; The National Institute of Standards and Technology (NIST, formerly known as the U.S. National Bureau of Standards); other governmental agencies (including all branches of the U.S. Military); industry trade associations (such as The Society of Automotive Engineers; NACE International; SSPC: The Society for Protective Coatings, formerly The Steel Structures Painting Council); major automobile manufacturing companies and their suppliers, and various test equipment manufacturers. Some improvements have included the establishment of required test specifications, along with tolerance limits for test variables, which help

produce more reliably reproducible and meaningful test results. Even with the newly revised test procedures and modern designs of salt spray test equipment, there are still test procedures, support information, variables, and specifications of the tests that need to be examined for inclusion or evolved to improve the test standards.

Reference cited in this section 1. W.D., McMaster, “A History of the Salt Spray Tests,” Publication, General Motors Corporation, Sept 1965

R. Singleton, Cabinet Testing, Corrosion: Fundamentals, Testing, and Protection ,Vol 13A, ASM Handbook, ASM International, 2003, p 470–477 Cabinet Testing Raymund Singleton, The Singleton Corporation

Applications and Use of Salt Spray (Fog) Testing The salt spray type test has received its widest acceptance as a tool for evaluating the uniformity of thickness and degree of porosity of metallic and nonmetallic protective coatings. It has served this purpose with a great deal of success. The test is useful for evaluating different lots of the same product once a standard level of performance has been established. It is especially helpful as a screening test for revealing a particularly inferior coating. In recent years, certain cyclic acidified salt spray tests have been used to test the resistance of aluminum alloys to exfoliation corrosion. The salt spray test is, also, considered very useful as an accelerated laboratory corrosion test that attempts to simulate the effects of marine atmospheres on different metals, with or without protective coatings. The most commonly used and accepted salt spray test methods in the United States and much of the rest of the world are the methods outlined in ASTM Standards B 117, B 368, and G 85 (Ref 2, 3, 4). U.S. Government agencies, professional associations, trade associations and groups, many industry groups, and automotive manufacturers and suppliers have established their own salt spray standards and procedures, but in the interest of utilizing existing successful technology, many of these standards largely conform to the details of the ASTM tests. Representative of these similar salt spray tests are GM 4298/Salt Fog and Mil Standard 810E, method 509.3. In addition, some international standards of similar type are DIN 50018 and ISO 9227, which includes salt fog, acetic-acid and copper accelerated salt spray (CASS) procedures. It is important to note that it is possible to perform widely utilized humidity corrosion test standard procedures of similar type and operation to the “salt-fog” tests in the same type test apparatus when properly equipped. The humidity tests, however, use deionized water as the only corrosive agent and do not include a “salt- type,” corrosion-accelerator electrolyte in the test solution. Examples of domestic and international “humidity” corrosion tests are ASTM D 1735 (humidity-fog type)(Ref 5), ASTM D 2247 (100% condensing humidity) (Ref 6), GM 4465 (humidity-fog type), and ASTM G 60 (cyclic humidity) (Ref 7). See Table 1.

Table 1 Commonly used cabinet corrosion test methods Test method

Static tests Salt fog Salt fog

Neutral salt fog

Salt fog

Salt fog

Acetic-acid salt fog Acetic-acid salt fog

Acetic-acid salt fog



Designation

Electrolyte solution Content Specific gravity

Step

Step title

Temperature °C °F

Relative humidity

Duration

pH

Comments

number

1.0255– 1.0400 1.0255– 1.0400

6.5– 7.2(a) 6.5– 7.2(a)

…

Salt fog

35

95

N/A

24 h/cycle(b) …

…

Salt fog

35

95

N/A

24 h/cycle(b) …

5% NaCl

1.029– 1.036

6.5–7.2

…

Salt fog

35

95

N/A

24 h/cycle(b) …

5% NaCl

1.0255– 1.0400

6.5– 7.2(a)

…

Salt fog

35

95

N/A

24 h/cycle(b) …

5% NaCl

1.0255– 1.0400

6.5– 7.2(a)

…

Salt fog

35

95

N/A

24 h/cycle(b) …

5% NaCl

1.0255– 1.0400 1.0255– 1.0400

3.1 to 3.3(c) 3.1 to 3.3(c)

…

Salt fog

35

95

N/A

144–240 h

…

…

Salt fog

35

95

N/A

TBD

…

5% NaCl

1.0255– 1.0400

3.1 to 3.3(c)

…

Salt fog

35

95

N/A

TBD

…

5% NaCl

1.0255– 1.0400

3.1 to 3.3(c)

…

Salt fog

35

95

N/A

TBD

…

5% NaCl

1.029– 1.036

3.1 to 3.3(c)

…

Salt fog

35

95

N/A

TBD

…

ASTM B 117

5% NaCl

DIN 50021-SS (Similar to ASTM B 117) ISO 9227-NSS (Similar to ASTM B 117) GM 4298P (Similar to ASTM B 117) Mil Std 810 E (Similar to ASTM B 117) ASTM G 85 A1 DIN 50021ESS (Similar to ASTM G 85 A1) DIN 50021ESS (Similar to ASTM G 85 A1) DIN 50021ESS (Similar to ASTM G 85 A1) ISO 9227AASS (Similar to ASTM G 85 A1)

5% NaCl

5% NaCl

Test method

Designation

Copper-accelerated ASTM B 368 acetic acid salt fog (CASS Test) Copper-accelerated DIN 50021acetic acid salt fog CASS (Similar to ASTM B 368) Copper-accelerated ISO 9227acetic acid salt fog CASS (Similar to ASTM B 368) ASTM D 1735 Humidity fog

Humidity fog

High humidity Corrodkote

Static or cyclic test Moist SO2

Cyclic tests Moist SO2

Electrolyte solution Content Specific gravity 5% NaCl 0.25 g 1.030– CuCl/L solution 1.040

Step

Step title

Temperature °C °F

Relative humidity

Duration

Comments

number … Salt fog

49

120

N/A

TBD

…

pH 3.1– 3.3(a)(c)

5% NaCl 0.25 g 1.030– CuCl/L solution 1.040

3.1– 3.3(a)(c)

…

Salt fog

50

122

N/A

TBD

…

5% NaCl 0.25 g 1.029– CuCl/L solution 1.036

3.1– 3.3(a)(c)

…

Salt fog

49

120

N/A

TBD

…

ASTM type IV water only

…

…

…

Humidity fog

38

100

N/A

24 h/cycle (com.)

GM 4465P ASTM type IV (Similar to water only ASTM D 1735) ASTM D 2247 ASTM type IV water only ASTM B 380 Special slurry(d) (ASTM D 2247)

…

…

…

Humidity fog

38

100

N/A

(b)

Similar to ASTM B 117, except without the salt …

…

…

…

Soak

38

100

100%

N/A

N/A

1 2

Slurry application Humidity fog

Ambient 38

Use air distributor pipe or equivalent. …

100

>50% 80–90%

Per customer 1 h (to dry) 20 h = 1 cycle

ASTM G 87 (Kesternich)

Deionized water inside the cabinet for wet bottom is the only solution (ASTM D 2247)

…

1

SO2 gas injection

Ambient

…

2

Soak(e)

40

104

N/A

3

Ambient(f) (method B only)

20–30

68–86

>75%

0.2–2 L as required Method A, 24 h; Method B, 8h 16 h

Deionized water inside the

…

1

SO2 gas injection

Ambient

0.2 or 2 L as specified

…

DIN 50018 (Kesternich)

…

…

Ambient

Ambient

Test method

Cyclic acidified acetic salt fog

Designation

(Similar to ASTM G 87, method B) ASTM G 85, A2

Acidified synthetic sea water (SWAAT test)

ASTM G 85, A3

Salt fog/SO2

ASTM G 85, A4

Electrolyte solution Content Specific gravity cabinet for wet bottom is the only solution 5% NaCl 1.0255– 1.0400

Synthetic sea water with 10 mL glacial acetic acid/L 5% NaCl or synthetic sea salt

1.0255– 1.0400

1.0255– 1.0400

Step

Temperature °C °F

Relative humidity

Duration

number 2 Soak(e) 3 Ambient(f) (B only)

40 20–30

104 68–86

N/A >75%

8h 16 h

1

Salt fog

49

120

¾h

…

2 3 1

Dry Soak Salt fog

…

Soak

1

Salt fog

120 120 75–120 (TBD) 75–120 (TBD) 95

2h 3¼h ½h

2

49 49 24–49 (TBD) 24–49 (TBD) 35

N/A (wet bottom) 7–40% 65–95% N/A (wet bottom) 75% N/A

b. 2 h(h) 1h 1h

pH

2.8 to 3.0(g)

2.8 to 3.0

2.5– 3.2(a)

2

3

Dilute electrolyte fog/dry

ASTM G 85, A5

0.05% NaCl 0.35% (NH4)2SO4

N/A

Scab corrosion creepback

GM 9511P

5% NaCl

N/A

Cyclic corrosion test

CCT IV

5% NaCl

N/A

Step title

SO2 addition

Soak

35

35

95

95

5.0– 5.4(a)

1 2

Salt fog Dry

Ambient 35

N/A

1 2

Dry/immersion Cold/ambient

3

Immersion/humidity

60/ambient 140/ambient N/A -13/ambient N/A 25/ambient Ambient/60 Ambient/140 N/A/85%

4 5 1 2 3

Ambient Humidity Salt fog Dry Humidity fog

N/A 60 50 60 60

Not specified

95

N/A 140 122 140 140

75 min 85% N/A 0. Independent measurements for metals such as zinc and copper (Ref 2, 3, 4, 6, 12, 13) that have corrosion products soluble in precipitation show that the cumulative precipitation runoff loss, MP(t), is essentially linear with respect to time, regardless of daily and seasonal variations in meteorology and air chemistry (Fig. 7). In other words, the precipitation runoff loss rate, dMP(t)/ dt, can be treated as a constant for a given geographic location or exposure site. Substituting Eq 3 with n = 1 into Eq 2 and integrating gives: MT(t) = (α/β)[1 - exp{-βMT(t)/α}] + βt

(Eq 6)

where β is the precipitation runoff loss rate, and α is a constant related to properties of the corrosion film.

Fig. 7 Cumulative copper precipitation runoff losses at marine unpolluted and polluted sites as a function of exposure time This empirical model describes the time variation of the total corrosion loss when the precipitation runoff loss rate is nonzero. It describes conditions where there is competition between corrosion film growth and corrosion film dissolution and removal in precipitation runoff. The precipitation runoff loss rate, β, describes the influence of environmental factors on corrosion film stability. The constant α describes the effect of intrinsic properties of minerals in the corrosion film on corrosion film stability. The ratio of the two terms, (α/β), describes the balance between the tendency for protective corrosion film growth and the environmental factors producing corrosion product losses in precipitation runoff. Equation 6 reduces to the parabolic form of Eq 4 in the limit where the precipitation runoff loss rate is 0. With sufficient environmental data and limited geochemical modeling, contributions of environmental processes to the precipitation runoff rate can be estimated. This has been done for zinc (Ref 12) and copper (Ref 13) exposed in a polluted urban environment, a marine environment, and a rural environment. Figure 8 shows four contributions to copper corrosion in the polluted urban environment: (a) a strong acid contribution from wet deposition representing the effect of acid rain, (b) a weak acid contribution in wet deposition representing the effect of the carbon dioxide/bicarbonate/carbonate equilibia, (c) a dry deposition contribution representing

acid gases (sulfur dioxide, nitric acid, hydrochloric acid) and perhaps modified by the deposition of basic particulates, and (d) a precipitation of brochantite, CuSO4 · 3Cu(OH)2, within the corrosion film as a result of weathering of the corrosion film. These contributions add together stoichiometrically to give the curve for the cumulative precipitation runoff loss, MP(t) = tβ.

Fig. 8 Dry deposition, strong acid, weak acid, and brochantite deposition contributions to cumulative copper precipitation runoff loss as determined by geochemical modeling for the polluted Washington, D.C. site Very long-term corrosion data that include both mass loss and corrosion film thickness and chemistry are not often available to test Eq 6. Data were available for a 100 year old copper roof in Stockholm, Sweden (Ref 5). Values of α and β were estimated, knowing that the ratio of the two values was equal to the steady-state thickness of the corrosion film. For an effective corrosion film thickness of 15.5 μm, the best estimates of the two values were (Ref 13) α = 13.9 μm2/year and β = 0.9 μm/year. The resulting curves for cumulative precipitation runoff loss, total corrosion mass loss, and the patina thickness are shown in Fig. 9 (Ref 13). This figure illustrates that corrosion film growth is the dominant process during the early exposure period. However, in long exposures, such as the 100 year old copper roof, a large fraction of the corrosion mass loss simply maintains the patina thickness so as to balance precipitation runoff losses. Of course, this is the result expected from Eq 2 for steady-state conditions where film growth has ceased, that is, dMF(t)/dt = 0. In that case, the corrosion rate equals exactly the precipitation runoff loss rate.

Fig. 9 Cumulative corrosion mass loss, corrosion film growth (patina), and precipitation runoff loss for a 100 year old copper roof exposed in Stockholm, Sweden Finally, the parameters in Eq 6, α and β, can be used to define a dimensionless corrosion equation, that is: MT(t*)* = [1 - exp{-MT(t*)*}] + t*

(Eq 7)

where MT(t*)* = MT(t)β/α, dimensionless cumulative corrosion loss; t* = tβ2/α, dimensionless time; MF,S = α/β, the steady-state corrosion film or patina thickness; tc = α/β2, a characteristic time (17.2 years for the example of the 100 year old copper roof). The dimensionless corrosion equation gives the curve shown in Fig. 10 as a function of time (Ref 13). Early in the corrosion process, the corrosion kinetics are parabolic, and the slope of the curve is . As the corrosion film matures and the corrosion process approaches steady state, the corrosion kinetics transition from parabolic kinetics to linear kinetics, and the slope changes from to 1. In the framework of the dimensionless equation, short and long exposures are relative terms defined by the balance of environmental factors (β) and factors related to the mineralogy of the corrosion film (α).

Fig. 10 Dimensionless empirical model for atmospheric corrosion with precipitation runoff losses showing effect of exposure time on cumulative mass loss and corrosion kinetics

References cited in this section 2. W. He, I. Odenvall Wallinder, and C. Leygraf, A Comparison Between Corrosion Rates and Runoff Rates from New and Aged Copper and Zinc as Roofing Materials, Water, Air Soil Pollut. Focus, Vol 1, 2001, p 67–82 3. I. Odnevall Wallinder, P. Verbiest, W. He, and C. Leygraf, The Influence of Patina Age and Pollutant Levels on the Runoff Rate of Zinc from Roofing Materials, Corros. Sci., Vol 40 (No. 11), 1998, p 1977–1982 4. I. Odnevall Wallinder, P. Verbiest, W. He, C. Leygraf, Effects of Exposure Direction and Inclination on the Runoff Rates of Zinc and Copper Roofs, Corros. Sci., Vol 42, 2000 p 1471–1487 5. W. He, I. Odnevall Wallinder, and C. Leygraf, A Laboratory Study of Copper and Zinc Runoff during First Flush and Steady- State Conditions, Corros. Sci., Vol 43, 2001, p 127–146 6. W. He, I. Odnevall Wallinder, and C. Leygraf, Runoff Rates of Zinc—A Four-Year Field and Laboratory Study, Outdoor Atmospheric Corrosion, STP 1421, H.E. Townsend, Ed., ASTM International, 2002, p 216–229 12. S.D. Cramer, S.A. Matthes, G.R. Holcomb, B.S. Covino, Jr., and S.J. Bullard, “Precipitation Runoff and Atmospheric Corrosion,” Paper 00452, Corrosion/2000, NACE International, 2000 13. S.D. Cramer, S.A. Matthes, B.S. Covino, Jr., S.J. Bullard, and G.R. Holcomb, Environmental Factors Affecting the Atmospheric Corrosion of Copper, Outdoor Atmospheric Corrosion, STP 1421, H.E. Townsend, Ed., American Society for Testing and Materials, 2002, p 245–264

48. R.H. McCuen, P. Albrecht, and J. Cheng, A New Approach to Power-Model Regression of Corrosion Penetration Data, Corrosion Forms and Control for Infrastructure, STP 1137, V. Chaker, Ed., American Society for Testing and Materials, 1992, p 46–76

S.D. Cramer and S.A. Matthes, Simulated Service Testing in the Atmosphere, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 487–494 Simulated Service Testing in the Atmosphere Stephen D. Cramer and Steven A. Matthes, Albany Research Center, U.S. Department of Energy

Acknowledgment Portions of this article have been adapted from K.L. Money, Simulated Service Testing: Corrosion Testing in the Atmosphere, Corrosion, Vol 13, Metals Handbook, 9th ed., ASM International, 1987, p 205–206.


References 1. Effects of Acidic Deposition on Materials, Report 19, Terrestrial, Materials, Health and Visibility Effects, Vol III, Acidic Deposition: State of Science and Technology, U.S. National Acid Precipitation Assessment Program, Washington, D.C., p. 19–1 to 19–280, 1991 2. W. He, I. Odenvall Wallinder, and C. Leygraf, A Comparison Between Corrosion Rates and Runoff Rates from New and Aged Copper and Zinc as Roofing Materials, Water, Air Soil Pollut. Focus, Vol 1, 2001, p 67–82 3. I. Odnevall Wallinder, P. Verbiest, W. He, and C. Leygraf, The Influence of Patina Age and Pollutant Levels on the Runoff Rate of Zinc from Roofing Materials, Corros. Sci., Vol 40 (No. 11), 1998, p 1977–1982 4. I. Odnevall Wallinder, P. Verbiest, W. He, C. Leygraf, Effects of Exposure Direction and Inclination on the Runoff Rates of Zinc and Copper Roofs, Corros. Sci., Vol 42, 2000 p 1471–1487 5. W. He, I. Odnevall Wallinder, and C. Leygraf, A Laboratory Study of Copper and Zinc Runoff during First Flush and Steady- State Conditions, Corros. Sci., Vol 43, 2001, p 127–146 6. W. He, I. Odnevall Wallinder, and C. Leygraf, Runoff Rates of Zinc—A Four-Year Field and Laboratory Study, Outdoor Atmospheric Corrosion, STP 1421, H.E. Townsend, Ed., ASTM International, 2002, p 216–229

7. X. Zhang, W. He, I. Odnevall Wallinder, J. Pan, and C. Leygraf, Determination of Instantaneous Corrosion Rates and Runoff Rates of Copper from Naturally Patinated Copper during Continuous Rain Events, Corros. Sci., Vol 44, 2002, p 2131–2151 8. I. Odnevall Wallinder and C. Leygraf, A Study of Copper Runoff in an Urban Atmosphere, Corros. Sci., Vol 39 (No. 12), 1997, p 2039–2052 9. S.D. Cramer and L.G. McDonald, Atmospheric Factors Affecting the Corrosion of Zinc, Galvanized Steel, and Copper, Corrosion Testing and Evaluation, STP 1000, R. Baboian and S.W. Dean, Ed., American Society for Testing and Materials, 1990, p 208–224 10. J.W. Spence, F.H Haynie, F.W. Lipfert, S.D. Cramer, and L.G. McDonald, Atmospheric Corrosion Model for Galvanized Steel Structures, Corrosion, Vol 48 (No. 12), 1992, p 1009–1019 11. S.D. Cramer, L.G. McDonald, and J.W. Spence, Effects of Acidic Deposition on the Corrosion of Zinc and Copper, Proceedings of the 12th International Corrosion Congress, Vol 2, NACE International, 1993, p 722–733 12. S.D. Cramer, S.A. Matthes, G.R. Holcomb, B.S. Covino, Jr., and S.J. Bullard, “Precipitation Runoff and Atmospheric Corrosion,” Paper 00452, Corrosion/2000, NACE International, 2000 13. S.D. Cramer, S.A. Matthes, B.S. Covino, Jr., S.J. Bullard, and G.R. Holcomb, Environmental Factors Affecting the Atmospheric Corrosion of Copper, Outdoor Atmospheric Corrosion, STP 1421, H.E. Townsend, Ed., American Society for Testing and Materials, 2002, p 245–264 14. S.A. Matthes, S.D. Cramer, B.S. Covino, Jr., S.J. Bullard, and G.R. Holcomb, Precipitation Runoff from Lead, Outdoor Atmospheric Corrosion, STP 1421, H.E. Townsend, Ed., American Society for Testing and Materials, 2002, p 265–274 15. S.A. Matthes, S.D. Cramer, S.J. Bullard, B.S. Covino, Jr., and G.R. Holcomb, “Atmospheric Corrosion and Precipitation Runoff from Zinc and Zinc Alloy Surfaces,” Paper 03598, Corrosion/2003, 17–20 March 2003 (San Diego, CA), NACE International, 2003 16. H.T. Michels, B. Boulanger, and N.P. Nikolaidis, “Copper Roof Stormwater Runoff—Corrosion and the Environment,” Paper 02225, Corrosion/2002, NACE International, 2002 17. H.T. Michels, B. Boulanger, and N.P. Nikolaidis, Environmental Impact of Stormwater Runoff from a Copper Roof, Mater. Perform., Vol 42 (No. 2) Feb 2003, p 70–74 18. M.C. Gromaire-Mertz, S. Garnaud, A. Gonzales, and G. Chebbo, Characterization of Urban Runoff Pollution in Paris, Wat. Sci. Technol., Vol 39 (No. 2), 1999, p 1–8 19. T.S. Barron, “Architectural Uses of Copper—An Evaluation of Stormwater Pollution Loads and Best Management Practices,” Palo Alto Regional Water Quality Control Plant, Palo Alto, CA, 2000 20. I. Odnevall Wallinder, C. Leygraf, C. Karlen, D. Heijerick, and C.R. Janssen, Atmospheric Corrosion of Zinc-Based Materials: Runoff Rates, Chemical Speciation and Ecotoxicity, Corros. Sci., Vol 43, 2001, p 809–816 21. “Notice of Proposed Standard for Sulfur Oxide, Particulate Matter, Carbon Monoxide, Photochemical Oxidants, Hydrocarbons, and Nitrogen Oxides,” 42 CFR, Part 410, Federal Register 36, Environmental Protection Agency, 1971, p 1502

22. F. Mansfield, “Regional Air Pollution Study Effects of Airborne Sulfur Pollutants on Materials,” EPA 600/4-80-007, Environmental Protection Agency, Jan 1980 23. F.H. Haynie and J.B. Upham, Effects of Atmospheric Sulfur Dioxide on the Corrosion of Zinc, Mater. Perform., Vol 9, 1970, p 35–40 24. H. Guttman and P.J. Sereda, Measurement of Atmospheric Factors Affecting the Corrosion of Metals, Metal Corrosion in the Atmosphere, STP 345, American Society for Testing and Materials, 1968, p 355–360 25. K.G. Compton, Atmospheric Corrosion, NACE Basic Corrosion Course, National Association of Corrosion Engineers, 1970, p 4-2 26. F.L. LaQue, Environment Factors in the Corrosion of Metals in Seawater and Sea Air, Marine Corrosion, John Wiley & Sons, 1975, p 103 27. D. Knotkova, V. Kucera, S.W. Dean, and P. Boschek, Classification of the Corrosivity of the Atmosphere—Standardized Classification System and Approach for Adjustment, Outdoor Atmospheric Corrosion, STP 1421, H.E. Townsend, Ed., American Society for Testing and Materials, 2002, p 109– 124 28. J. Tidblad, V. Kucera, A.A. Mikhailov, and D. Knotkova, Improvement of the ISO Classification System Based on Dose-Response Functions Describing the Corrosivity of Outdoor Atmospheres, Outdoor Atmospheric Corrosion, STP 1421, H.E. Townsend, Ed., American Society for Testing and Materials, 2002, p 73–87 29. M. Morcillo, E. Almeida, B. Chico, and D. de la Fuente, Analysis of ISO Standard 9223, Outdoor Atmospheric Corrosion, STP 1421, H.E. Townsend, Ed., American Society for Testing and Materials, 2002, p 59–72 30. C. Leygraf and T.E. Graedel, Atmospheric Corrosion, Wiley-Interscience, 2000, p 109–115 31. F.L. LaQue, Proc. ASTM, Vol 51, 1951, p 495 32. J.C. Hudson, J. Br. Iron Steel Inst., Vol 148, 1943, p 161 33. S.K. Coburn et al., Corrosiveness of Various Atmospheric Test Sites as Measured by Specimens of Steel and Zinc, Metal Corrosion in the Atmosphere, STP 435, American Society for Testing and Materials, 1968, p 326 34. ASTM Mater. Res. Stand., Dec 1961, p 977 35. S.W. Dean, Jr., Planning, Instrumentation and Evaluation of Atmospheric Corrosion Tests and a Review of ASTM Testing, Atmospheric Corrosion, John Wiley & Sons, 1982, p 195–216 36. “Standard Recommended Practice for Conducting Atmospheric Corrosion Tests on Metals,” G 50, Annual Book of ASTM Standards, American Society for Testing and Materials 37. “Standard Recommended Practice for Making and Using U-Bend Stress Corrosion Test Specimens,” G 30, Annual Book of ASTM Standards, American Society for Testing and Materials 38. “Standard Practice for Preparation and Use of Bent Beam Stress Corrosion Test Specimens,” G 39, Annual Book of ASTM Standards, American Society for Testing and Materials

39. “Standard Recommended Practice for Making and Using C-Ring Stress Corrosion Test Specimens,” G 38, Annual Book of ASTM Standards, American Society for Testing and Materials 40. “Standard Recommended Practice for Preparation and Use of Direct Tension Stress Corrosion Test Specimens,” G 49, Annual Book of ASTM Standards, American Society for Testing and Materials 41. “Standard Practice for Preparation of Stress Corrosion Test Specimens for Weldments,” G 58, Annual Book of ASTM Standards, American Society for Testing and Materials 42. R. Baboian, Final Report on the ASTM Study: Atmospheric Galvanic Corrosion of Magnesium Coupled to Other Metals, Atmospheric Factors Affecting the Corrosion of Engineering Metals, STP 646, S.K. Coburn, Ed., American Society for Testing and Materials, 1978, p 17–29 43. “Metals and Alloys—Test Methods for Contact Corrosion in Atmospheric Conditions,” ISO/TC 156/WG 3/N11, Third Working Draft, Proposal of USSR to ISO Technical Committee 156 WG 3, Nov 1979 44. “Corrosion of Metals and Alloys—Corrosivity of Atmospheres—Methods of Determination of Corrosion Rates of Standard Specimens for the Evaluation of Corrosivity,” ISO/TC 156 WG4/N131, German Proposal to ISO Technical Committee 156 WG3, 1979 45. T.E. Graedel, K. Nassau, and J.P. Franey, Copper Patinas Formed in the Atmosphere, Part I: Introduction, Corros. Sci., Vol 27 (No. 7), 1987, p 639–657 46. T.E. Graedel, Copper Patinas Formed in the Atmosphere, Part II: Qualitative Assessment of Mechanisms, Corros. Sci., Vol 27 (No. 7), 1987, p 721–740 47. T.E. Graedel, Copper Patinas Formed in the Atmosphere, Part III: A Semi-Quantitative Assessment of Rates and Constraints in the Greater New York Metropolitan Area, Corros. Sci., Vol 27 (No. 7), 1987, p 741–769 48. R.H. McCuen, P. Albrecht, and J. Cheng, A New Approach to Power-Model Regression of Corrosion Penetration Data, Corrosion Forms and Control for Infrastructure, STP 1137, V. Chaker, Ed., American Society for Testing and Materials, 1992, p 46–76


Selected References Special Technical Publications from ASTM International: • • •

R. Baboian and S.W. Dean, Jr., Ed., Corrosion Testing and Evaluation, STP 1000, 1990 V. Chaker, Ed., Corrosion Forms and Control for Infrastructure, STP 1137, 1992 S.K. Coburn, Ed., Atmospheric Factors Affecting the Corrosion of Engineering Materials, STP 646, 1978

• • • • • • •

Corrosion in Natural Environments, STP 558, 1974 S.W. Dean, Jr., G.H. Delgadillo, and J.B. Bushman, Ed., Marine Corrosion in Tropical Environments, STP 1399, 2000 S.W. Dean, Jr. and T.S. Lee, Ed., Degradation of Metals in the Atmosphere, STP 965, 1988 S.W. Dean, Jr. and E.C. Rhea, Ed., Atmospheric Corrosion of Metals, STP 767, 1982 W.W. Kirk and H.H. Lawson, Ed., Atmospheric Corrosion, STP 1239, 1995 Metal Corrosion in the Atmosphere, STP 435, 1968 H.E. Townsend, Ed., Outdoor Atmospheric Corrosion, STP 1421, 2002

Publications from other sources: • • • • •

W.H. Ailor, Ed., Atmospheric Corrosion, Wiley-Interscience, 1982 R. Baboian, Ed., Materials Degradation Caused by Acid Rain, ACS Symposium Series 318, American Chemical Society, 1986 V. Kucera and E. Mattsson, Atmospheric Corrosion, Corrosion Mechanisms, F. Mansfeld, Ed., Marcel Dekker, 1987, p 211–284 C. Leygraf and T.E. Graedel, Atmospheric Corrosion, Wiley-Interscience, 2000 X.G. Zhang, Corrosion and Electrochemistry of Zinc, Plenum Press, 1996

Simulated Service Corrosion Testing in Water, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 495–496

Simulated Service Testing in Water Introduction SIMULATED SERVICE TESTING is the most reliable predictor of corrosion behavior short of actual service experience. This includes exposures of either structural components or test specimens in environments that are representative of many general service situations. Test materials are subjected to the cyclic effects of the weather, geographical influences, and bacteriological factors that cannot be realistically duplicated in the laboratory. This type of testing is important for such objectives as materials selection, predicting the probable service life of a product or structure, evaluating new commercial alloys and processes, and calibrating laboratory corrosion tests. The type of information sought determines the selection of test specimens and the methods of assessing the corrosion effects. As an environment in which corrosion occurs, water can be categorized as treated or natural. Regardless of the original intent for which a test has been developed, application can be extended to include any use. For example, ASTM G 78 (described in Table 1) is a guide for crevice corrosion testing of iron- and nickel-base alloys in seawater. The crevice assembly described could be installed on any alloy type in any environment if crevices are being studied. Only when quality control or acceptance stipulates a particular test method is there no room for departure. Some test methods involve the insertion of coupons or test panels into a conduit, with the environment at actual velocity and aeration. Other test methods involve the insertion of a test section into the flow line, especially if wall effects might be important. Still other tests require immersion of specimens in relatively open bodies of water.

Table 1 ASTM International and NACE methods for testing waters and materials in waters Designation Title ASTM standards Standard Test Methods for pH of D 1293 Water

D 1498

Standard Practice for OxidationReduction Potential of Water

D 2688

Standard Test Methods for Corrosivity of Water in the Absence of Heat Transfer (Weight Loss Methods)

E 645

Standard Test Method for Efficacy of Microbiocides Used in Cooling Systems

G 52

Standard Practice for Exposing and Evaluating Metals and Alloys in Surface Seawater

G 78

Standard Guide for Crevice Corrosion Testing of Iron-Base and Nickel-Base Stainless Alloys in Seawater and Other ChlorideContaining Aqueous Environments

NACE standards Laboratory Methods for the TM 0174 Evaluation of Protective Coatings Used as Lining Materials in Immersion Service

RP 0775

Preparation and Installation of Corrosion Coupons and Interpretation of Test Data in Oil

Description These methods cover the determination of pH by electrometric measurement, using the glass electrode as the sensor. Two procedures are discussed. Method A covers the precise laboratory measurement of pH in water with the use of at least two of seven standard reference buffer solutions for equipment standardization. Method B concerns the routine or continuous measurement of pH in the laboratory and the measurement of pH under various continuous process conditions. This practice covers the apparatus and procedure for the electrometric measurement of oxidation-reduction potential in water. These methods cover the determination of the corrosivity of water by evaluating pitting and by measuring the weight loss of metallic specimens. Additional information can be found in the articles “Pitting Corrosion” and “Evaluating Pitting Corrosion” in this Volume. This method covers the effect of bactericides and fungicides in controlling microbial growth in cooling water by using water collected from operational cooling water systems. Additional information can be found in the articles “Microbiologically Influenced Corrosion” and “Evaluating Microbiologically Influenced Corrosion” in this Volume. This practice covers conditions for exposure of metals and alloys to surface seawater and presents the general procedures that should be followed in conducting seawater exposure tests so that meaningful comparisons can be made for different locations. This guide provides information for conducting crevice corrosion tests and identifies factors that may affect results and influence conclusions. These procedures can be used to identify the conditions that will most likely result in crevice corrosion and to provide a basis for assessing the relative resistance of various alloys to crevice corrosion under certain specified conditions. Additional information on crevice corrosion can be found in the articles “Crevice Corrosion” and “Evaluating Crevice Corrosion” in this Volume. This standard gives guidelines to assist manufacturers and users of protective coatings in selecting materials by providing standard test methods for evaluating protective coatings used as linings for immersion service. Two test methods are discussed for the evaluation of protective coatings on any substrate, such as steel, copper, or aluminum, so that both the factors of chemical resistance and permeability can be considered. This standard emphasizes the use of uniform industryproven methods to monitor corrosion in oil production systems. Procedures for preparing, analyzing, and installing

Production Practice corrosion coupons are also outlined. Tests used as criteria for acceptance and/or quality control may require a very specific water chemistry and values of such factors as hardness, pH, aeration, temperature, alkalinity, and specific dissolved ion concentrations. It is usually valuable to monitor some of these factors in testing any water because compositions and contaminants change with time as sources (well, river, estuary, storm drains) and climatic conditions (wind direction, temperature, acid rain) vary. A variation in the mixture of sources might change test results significantly. It is important to be aware of the factors that can affect test results when interpreting those results. Unless galvanic action is being investigated (and it probably should be in most cases), specimens of differing alloy composition or thermomechanical treatment should be electrically isolated or insulated from each other and from the support framework. The factors that require consideration in water supplies are discussed in Ref 1, 2, 3, 4, 5. Many standardized and nonstandardized tests, methods, and practices are discussed in Ref 6, 7, 8, 9, 10. The ASTM International and NACE International standards that are directly or indirectly applicable to simulated service corrosion testing in water are listed and described in Table 1.

References cited in this section 1. F.N. Speller, in Corrosion Handbook, H. Uhlig, Ed., John Wiley & Sons, 1948, p 496 2. F.N. Speller, Corrosion: Causes and Prevention, McGraw-Hill, 1951, p 244 3. R.K. Swandby, Corrosion Resistance of Metals and Alloys, 2nd ed., F.L. LaQue and H.R. Copson, Ed., Reinhold, 1963, p 45 4. G. Butler and H.C.K. Ison, Corrosion and Its Prevention in Waters, Reinhold, 1966, p 18 5. Prevention and Control of Water-Caused Problems in Building Potable Water Systems, TPC 7, National Association of Corrosion Engineers, 1980 6. V.V. Kendall, in Corrosion Handbook, H. Uhlig, Ed., John Wiley & Sons, 1948, p 1058 7. F.L. LaQue, in Corrosion Handbook, H. Uhlig, Ed., John Wiley & Sons, 1948, p 1060 8. F.L. LaQue, in Corrosion Resistance of Metals and Alloys, 2nd ed., F.L. LaQue and H.R. Copson, Ed., Reinhold, 1963, p 107, 136 9. O.B.K. Fraser, in Corrosion Handbook, H. Uhlig, Ed., John Wiley & Sons, 1948, p 105 10. F.A. Champion, Corrosion Testing Procedures, 2nd ed., John Wiley & Sons, 1965

Simulated Service Corrosion Testing in Water, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 495–496 Simulated Service Testing in Water

Test Method Selection and Precautions It is necessary to exercise sound judgment when selecting suitable testing procedures. For example, the results of static-immersion tests or alternate-immersion tests bear little relationship to pipelines flowing full within an operating plant. Test coupons inserted into an operating process line, or at least into a laboratory test loop using

process fluids at operating temperatures and velocities, will produce results that agree much better with actual plant experience. Because corrosion is often accelerated at metal/ liquid/vapor interfaces, these interfaces should be duplicated in the testing procedure if they are present in the plant or should be omitted from the testing if they are not an actual factor.


Test Specimens Specimen geometry, ratio of environment solution volume to exposed specimen area, or other shape and area factors should be as similar as possible to true plant conditions. Exposed, cut specimen edges, which are often attacked more rapidly than other parts, should be examined and discounted if edges are not exposed in plant equipment. Similarly, stamped specimen identification characters may suffer increased rates of attack or may be sites of stress-assisted or accelerated corrosion. Virtually every specimen size has been used in immersion testing. Popular laboratory test specimen sizes range from 1 cm2 (0.15 in.2) of exposed surface to 25 × 100 mm (1 × 4 in.), both of which are used frequently, to 100 × 150 mm (4 × 6 in.). In-plant and seawater tests sometimes involve specimens measuring 150 × 250 mm (6 × 10 in.) or larger. When weldments or projections are of concern, either the weldment or the projection should be incorporated into the specimen. Entire components are seldom tested in the laboratory but may be installed in pilot plant or works scale operations (pumps, condensers, heat exchangers).


Effect of Water Variables To date, no compilation of relative rankings of alloy types in various aqueous environments is available. However, for iron-base alloys, the corrosion rate is relatively uniform in most waters, whether they are of high or low purity. The rate seems to vary more with aeration, that is, from less than 0.05 mm/yr (50 mils/ yr) in aerated water. In tap, brackish, or seawater, pitting tends to occur. In tap water and seawater, yellow brass tends to dezincify (especially when conditions are near stagnation), although corrosion rates are less than 0.05 mm/yr (20 mils/yr) for tap water and seawater, respectively. Detailed information on dezincification can be found in the article “Effects of Metallurgical Variables on Dealloying Corrosion” in this Volume. There are no definite rules concerning the effect of flow rates, temperatures, and/or pH on the corrosion of alloys in water systems, although in waters free of passivators (including dissolved oxygen), increasing flow rates tend to erode any protective corrosion films that may have formed. Corrosion rates of most alloys over the pH range of 4 to 10 show little effect of pH. Outside this range, that is, 0 to 4 and 10 to 14, the environment cannot technically be considered water, except for some highly unusual acid rain cases. Increasing temperatures sometimes precipitate protective salts, such as calcium carbonate, which decrease corrosion rates in normal-tohard waters. Very pure or very soft waters are often excellent solvents for metallic ions, such as copper. If these waters are used to prepare laboratory solutions for test purposes, dissolved copper from copper water lines can deposit or

plate out on more active metal surfaces. In the case of aluminum alloys, this deposited copper can greatly accelerate corrosion immediately adjacent to the copper. The effect of dissolved copper ion on localized corrosion of aluminum is cited in ASTM Standards G 4, section 4.5 (Ref 11); G 52, section 6.2 (Ref 12); and G 71, section 4.2.1 (Ref 13). Similar effects of lesser intensity have been observed on most active metals. Temperature differentials between points in a flow system can produce accelerated attack due to differences in ionic activity. Although this attack usually occurs at the point of higher temperature, protective scales occasionally precipitate on the high-temperature surface, with attack occurring at the cooler sites. In any system involving differential temperature, the investigator should be aware of this possible behavior. Another temperature effect in open, aqueous systems is the decreasing solubility of gases, particularly oxygen, with increasing temperature. This effect reduces the cathodic action, or more exactly that portion due to oxygen reduction, and thus decreases the amount of anodic reaction that occurs. Similarly, the solubility of air and oxygen in saline solutions decreases with increasing concentration of salt, but the solution conductivity increases with the dissolved salt concentration. The two effects combine in oxygen reduction cathodic systems to produce increasing corrosion rates up to about 3.5 wt% sodium chloride (NaCl) solutions and decreasing corrosion rates above that value. This particular effect is not observed in freshwater systems. Natural waters, depending on the source, may contain variable concentrations of dissolved carbonates. These are principally calcium, magnesium, and/or sodium, with the solubility controlled by both the partial pressure of carbon dioxide and the solubility product of calcium carbonate as it varies with alkalinity and temperature. Deposition of a thin continuous film of carbonate can protect the underlying metal where a lack of the same would allow attack. On the other hand, a film that is too thick impedes heat transfer and can lead to reduced flow and ultimately to possible tube burn-out. Carbonate scaling, measurement through the Langelier Index, and controlling treatments are discussed in Ref 14.

References cited in this section 11. “Standard Guide for Conducting Corrosion Coupon Tests in Field Applications,” G 4, Annual Book of ASTM Standards, American Society for Testing and Materials 12. “Standard Practice for Exposing and Evaluating Metals and Alloys in Surface Seawater,” G 52, Annual Book of ASTM Standards, American Society for Testing and Materials 13. “Standard Guide for Conducting and Evaluating Galvanic Corrosion Tests in Electrolytes,” G 71, Annual Book of ASTM Standards, American Society for Testing and Materials 14. H.H. Uhlig and R.W. Revie, Corrosion and Corrosion Control, 3rd ed., John Wiley & Sons, 1985, p 111–114, 415–420


Acknowledgment This article has been adapted from C.H. Baloun, Corrosion Testing in Water, Corrosion, Vol 13, ASM Handbook (formerly Metals Handbook 9th ed.,), ASM International, 1987, pages 207–208.


References 1. F.N. Speller, in Corrosion Handbook, H. Uhlig, Ed., John Wiley & Sons, 1948, p 496 2. F.N. Speller, Corrosion: Causes and Prevention, McGraw-Hill, 1951, p 244 3. R.K. Swandby, Corrosion Resistance of Metals and Alloys, 2nd ed., F.L. LaQue and H.R. Copson, Ed., Reinhold, 1963, p 45 4. G. Butler and H.C.K. Ison, Corrosion and Its Prevention in Waters, Reinhold, 1966, p 18 5. Prevention and Control of Water-Caused Problems in Building Potable Water Systems, TPC 7, National Association of Corrosion Engineers, 1980 6. V.V. Kendall, in Corrosion Handbook, H. Uhlig, Ed., John Wiley & Sons, 1948, p 1058 7. F.L. LaQue, in Corrosion Handbook, H. Uhlig, Ed., John Wiley & Sons, 1948, p 1060 8. F.L. LaQue, in Corrosion Resistance of Metals and Alloys, 2nd ed., F.L. LaQue and H.R. Copson, Ed., Reinhold, 1963, p 107, 136 9. O.B.K. Fraser, in Corrosion Handbook, H. Uhlig, Ed., John Wiley & Sons, 1948, p 105 10. F.A. Champion, Corrosion Testing Procedures, 2nd ed., John Wiley & Sons, 1965 11. “Standard Guide for Conducting Corrosion Coupon Tests in Field Applications,” G 4, Annual Book of ASTM Standards, American Society for Testing and Materials 12. “Standard Practice for Exposing and Evaluating Metals and Alloys in Surface Seawater,” G 52, Annual Book of ASTM Standards, American Society for Testing and Materials 13. “Standard Guide for Conducting and Evaluating Galvanic Corrosion Tests in Electrolytes,” G 71, Annual Book of ASTM Standards, American Society for Testing and Materials 14. H.H. Uhlig and R.W. Revie, Corrosion and Corrosion Control, 3rd ed., John Wiley & Sons, 1985, p 111–114, 415–420


Selected References

• •

J.F. Jenkins, Seawater Tests, Corrosion Tests and Standards: Application and Interpretation, R. Baboian, Ed., ASTM International, 1995, p 127–130 W.T. Young and P. Fairer, Freshwater Tests, Corrosion Tests and Standards: Application and Interpretation, R. Baboian, Ed., ASTM International, 1995, p 131–136

M.K.A. Flitton, and E. Escalante, Simulated Service Testing in Soil, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 497–500

Simulated Service Testing in Soil M.K. Adler Flitton, Idaho National Engineering and Environmental Laboratory; E. Escalante, Corrosion Group, National Institute of Science and Technology (retired)

Introduction TEST METHODS for evaluating the durability of a metal in soil are described in this article. Specimen design, preparation, burial, and retrieval techniques are discussed. The type of information sought during soil-induced corrosion evaluation controls the design configuration and the nature of the corrosion measurements. Consideration of these factors during the planning stage will help the corrosion engineer obtain a maximum amount of information with a minimum number of problems. The corrosion of metals in soils, which is also called underground corrosion, can be divided into two broad categories: corrosion in undisturbed soils and corrosion in disturbed soils. Corrosion in undisturbed soil is always low, regardless of soil conditions, and is limited only by the availability of the oxygen necessary for the cathodic reaction. Steel piles driven into soil fall under this category and therefore undergo limited corrosive attack. Corrosion of metals in disturbed soils is strongly affected by soil conditions. Soil electrical resistivity, mineral composition, dissolved salts, moisture content, total acidity or alkalinity (pH), redox potentials, microbiological activity, and concentration of oxygen play important roles in determining whether metal corrosion is a serious problem in soils (Table 1). Other, secondary factors are also important but are more difficult to define. Thus, simply testing metals and alloys in a variable pH solution or in aerated or deaerated solutions will not accurately simulate the conditions in soil (Ref 2). Table 1 Numerical corrosivity scale Point system for predicting soil corrosivity according to American Water Works Association C 105 Soil parameter Value Assigned points 2000 0 0–2 5 pH 2–4 3 4–6.5 0 6.5–7.5 0 7.5–8.5 0 >8.5 3 0 Redox potential, mV >100

50–100 3.5 0–50 4 20,000 Mildly corrosive 10,000–20,000 Moderately corrosive 5,000–10,000 Corrosive 3,000–5,000 Highly corrosive 1,000–3,000 Extremely corrosive t. In this limit, the quantity

approaches 1, and the thermal contrast approaches a very simple relation ΔT = Q/ρc(1/d - 1/t). This relationship has been observed and reported in many scientific papers.

References cited in this section 6. I. Perez, R. Santos, P. Kulowitch, and M. Ryan, Calorimetric Modeling of Thermographic Data, in Thermosense XX, Society of Photo-Optical Instrumentation Engineers, Vol 3361, 1998, p 75

7. I. Perez, W.R. Davis, P. Kulowitch, and M.F. Dersch, Thermographic Modeling of Water Entrapment, Thermosense XXI, D.H. LeMieux and J.R. Snell, Jr., Society of Photo- Optical Instrumentation Engineers, April 1999, p 32–39 8. I. Perez and S. Shepard, Modeling of Pulsed Thermography in Anisotropic Media, Review of Progress in QNDE, Vol 18B, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1998, p 2063–2072

I. Perez, Infrared Imaging for Corrosion, Disbondments, and Cracks, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 525–528 Infrared Imaging for Corrosion, Disbondments, and Cracks Ignacio Perez, Office of Naval Research

Imaging Results The detection sensitivity of the pulse thermographic process, like ΔTmax, depends on the amount of energy deposited on the surface of the panel, the deposition time (Q, τdep), the size and depth of the flaw and thickness of the panel (R, d, t) (Ref 9). In general terms, the sensitivity increases linearly with the amount of energy, Q, and decreases inversely to the distance of the defect from the surface, d. Other factors that will significantly affect the sensitivity of the technique are edge effects, fasteners with different thermal conductivity than the base metal, and variations of an adhesive layer thickness in multilayered structures. Presently, the resolution achieved with the state-of-the-art system is about 3% thickness loss in laboratory environments inspecting flat panels and about 15 to 20% thickness loss under realistic field conditions inspecting structural components. Pulse thermographic methods work best when inspecting materials with defects that are parallel to the surface such as disbondments, delaminations, and generalized corrosion because they disrupt the flow of energy very effectively. When the defect is perpendicular to the surface of the sample, such as with many types of cracks, pulse thermography does not work since disruption to the energy flow is minimal. Other thermographic methods are being developed for these types of defects such as line-scanning thermography and, more recently, sonic thermography.

Reference cited in this section 9. I. Perez and X. Han, Pulsed Thermography Modeling, Review of Progress in QNDE, Vol 21, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 2002


Sonic Thermography for Crack Detection

Original work on sonic thermography started in the late 1970s with studies on the interaction between highpower ultrasonic sources and materials (Ref 10). It was found accidentally that when a cracked specimen was insonified with a high-power ultrasonic gun, cracks would be excited in such a way that they were easily detected with a thermal camera. Recent work in sonic thermography or thermosonics (Ref 11, 12) has demonstrated that this technique can also be applied to detect other forms of damage in materials, such as delamination in composites or corrosion in aluminum. In the thermosonic method, a burst of high- power ultrasonic energy is applied to a sample. The energy quickly floods all corners of the sample. If a crack is present anywhere on the sample, energy will dissipate very effectively from it, creating a region of elevated temperature that can be imaged effectively with an IR camera (Fig. 3).

Fig. 3 Sonic thermography arrangement. (a) Ultrasonic (UT) gun energizes sample. Cork insulators mechanically isolate samples from the backing plate. Infrared (IR) camera captures images. (b) Thermograph from IR camera detects a crack at the root of the slot.

This mode of coupling energy to a sample is much more effective than the one used in the pulse thermographic method. In the pulse thermographic method, energy must be deposited uniformly over the entire surface of the sample being inspected, but it is only the energy that falls over the defective area that contributes to the thermographic signal. The remainder of the incident energy contributes to the background temperature, thereby reducing the maximum possible thermal contrast, since the thermal contrast is obtained by subtracting the background temperature from the temperature of the hot spot. In the thermosonic method, energy is coupled to the sample being inspected with an ultrasonic (UT) gun in contact with the sample at one point of its surface. The energy then quickly spreads to all corners of the sample, with the amplitude of vibration building up in the sample while the UT gun is on. If there is a crack present in the sample, vibrational energy will start dissipating through it. The dissipation mechanism is not completely understood, but it is believed to be related to rubbing or clapping of the facing surfaces of the crack. The net effect of this dissipation is that the area surrounding the crack will heat up, thereby allowing for its detection. This is a very effective use of the interrogation energy, since most of it is consumed in generating the thermosonic signal. An analogy to detecting a crack by using the thermosonic process is finding a hole on a bicycle tire by using an air pressure gun. The analogy can be explained as follows. To find a hole in a tire, one applies air pressure through one point on the surface of the tire. The air quickly fills the entire volume of the tire and the pressure keeps building up in the tire while the pump remains on. If there is a small hole in the tire, air will leak through it, generating a small sound, thereby allowing for its detection. If there are several holes in the tire, each one will leak air, allowing for their detection. The higher the pressure in the tire, the stronger the sound produced by the hole. Continuing with the analogy, we know that a minimum amount of air pressure is required to detect a hole in a tire. Very small holes are effectively closed due to the elastic properties of rubber, and a minimum amount of air pressure is required to force air through the hole. Likewise in thermosonic experiments, a minimum amount of energy is required to excite cracks. This should be expected since closed cracks, such as hairline fatigue cracks that are almost invisible to the naked eye, would require a minimum amount of vibration amplitude to overcome the frictional forces produced by the mating surfaces. Evidence of this phenomena has been observed in laboratory experiments. Figure 3 shows the experimental setup used by the author. A 1000 W power supply (not shown in the figure) was used, and the excitation frequency was 20 kHz. It would be desirable to have a system where the frequency of the UT gun could be adjusted depending on the particular characteristics of the sample and defect being inspected. Unfortunately, a recent search found all high-power commercial units had a fixed excitation frequency. This is because these units are engineered to achieve maximum amplitude displacements at the tip of the gun by stacking up a series of precisely cut piezoelectric crystals operating at the same resonant frequency. A steel horn (Fig. 3) is used in the UT gun to enhance the amplitude of vibration. During a standard experiment, the controller is set to maximum power with pulse duration adjusted according to the experimental requirements, but limited to no more than 0.5 s because of hardware constraints. In fact, the longer the excitation, the stronger the thermosonic signal will be (Ref 13). The UT gun was mounted on a platform to impede it from sliding over the surface of the sample and damaging the surface, while providing good load control via a pneumatic regulator. The load applied to the gun against the sample determines the amount of energy that couples into the sample and thereby determines the amplitude of the thermosonic signal. It was found in experiments that the thermosonic signal increased linearly with the applied load. Conservation of energy dictates that at some higher load value the thermosonic signal will saturate, but the available load capacity of 45 lb was far from saturation. As high loads tend to scourge the surface of the sample, a scientist at Wayne State University placed a thin coupling agent (a thin foil of copper) between the UT gun and the sample to minimize surface damage while maintaining good mechanical coupling. The appropriate choice of the coupling agent will determine if this technique is truly nondestructive and how much energy couples into the part being inspected. The infrared camera used in all the experiments discussed in this paper was a Radiance HS camera with a frame rate adjusted to capture thermal images at a rate of 120 frames/s. Mechanical isolation of the sample is required in order to obtain an optimal thermosonic signal. If the sample being inspected rested directly over a backing plate of similar acoustic impedance or if the mechanical isolation between the sample and the backing plate was not significant, then the thermosonic signal would be minimal or not existing at all. This lack of thermosonic signal is understandable because the ultrasonic energy would leak out of the sample through the backing plate. Using the bicycle tire analogy, trying to detect a small fatigue crack in a sample in direct contact with a backing plate would correspond to trying to detect a small pinhole on

the tire when there was a large open gap in the tire. Air pressure would not be able to build in the tire due to the presence of the open gap in the tire. Similarly, good thermosonic signals are obtained when a piece of cork is placed between the sample and the backing plate to produce good mechanical isolation (Fig. 3) (Ref 13). In practice, complete isolation, while desirable, is difficult to achieve. This does not mean that thermosonics is limited to the inspection of well-isolated components. As long as the energy leakage to the surrounding area is small compared to the energy required to excite a crack, thermosonics will be useful as an NDE technique. There is still considerable experimental work and development to be done before thermosonics becomes a userfriendly NDE technique. The thermographic imaging techniques presented in this article and other techniques currently being developed are expanding the use of thermography into areas not previously considered.

References cited in this section 10. R.B. Mignogna, R.E. Green Jr., J.C. Duke Jr., E.G. Henneke II, and K.L. Reifsnider, Ultrasonics, Vol 19, 1981, p 159–163 11. L.D. Favro, X. Han, Z. Ouyang, G. Sun, H. Sui, and R.L. Thomas, Rev. Sci. Instrum., Vol 71, 2000, p 2418 12. J. Rantala, D. Wu, and G. Busse, Amplitude Modulated Lock-in Vibrothermography for NDE of Polymers and Composites, Research in Nondestructive Evaluation, Vol 7, No. 4, 1995, p 215–228 13. I. Perez and W.R. Davis, Optimizing the Thermosonic Signal, Review of Progress in QNDE, Vol 23, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 2003


Acknowledgment The author would like to thank Dr. Robert Green from The Johns Hopkins University and Dr. Skip Favro and Dr. Xiaoyan Han from Wayne State University for insightful observations and valuable critiques of this article. Thanks to Mr. Richard Wood, Naval Air warfare center at Cherry Point for assistance with figures.


References

1. H. Xiaoyan, L.D. Favro, and R.L. Thomas, Quantitative Defect Depth Measurements for NDE of Composites, Proc. American Society for Composites, R.F. Gibson and G.M. Newaz, Ed., Technomic Publishing Co., 1998, p 1077–1081 2. H. Xiaoyan, L.D. Favro, and R.L. Thomas, Thermal Wave NDI of Disbonds and Corrosion in Aircraft, Second Joint NASA/FAA/ DOD Conf. Aging Aircraft, NASA/CP- 1999-208982/Part 1, C.E. Harris, Ed., Langley Research Center, p 265–274 3. M.B. Saintey and D.P. Almond, Mathematical Modelling of Transient Thermography and Defect Sizing, Review of Progress in QNDE, Vol 15, D.O. Thompson and D.E. Chimenti, Ed., Plenum, 1996, p 503 4. L.D. Favro, X. Han, P.K. Kuo, and R.L. Thomas, Imaging the Early Time Behavior of Reflected Thermal Wave Pulses, in Thermosense XVII, Society of Photo-Optical Instrumentation Engineers, Vol 2473, 1995, p 162 5. X.P.V. Maldague, Nondestructive Evaluation of Materials by IR Thermography, Springer-Verlag, 1993, p 29 6. I. Perez, R. Santos, P. Kulowitch, and M. Ryan, Calorimetric Modeling of Thermographic Data, in Thermosense XX, Society of Photo-Optical Instrumentation Engineers, Vol 3361, 1998, p 75 7. I. Perez, W.R. Davis, P. Kulowitch, and M.F. Dersch, Thermographic Modeling of Water Entrapment, Thermosense XXI, D.H. LeMieux and J.R. Snell, Jr., Society of Photo- Optical Instrumentation Engineers, April 1999, p 32–39 8. I. Perez and S. Shepard, Modeling of Pulsed Thermography in Anisotropic Media, Review of Progress in QNDE, Vol 18B, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1998, p 2063–2072 9. I. Perez and X. Han, Pulsed Thermography Modeling, Review of Progress in QNDE, Vol 21, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 2002 10. R.B. Mignogna, R.E. Green Jr., J.C. Duke Jr., E.G. Henneke II, and K.L. Reifsnider, Ultrasonics, Vol 19, 1981, p 159–163 11. L.D. Favro, X. Han, Z. Ouyang, G. Sun, H. Sui, and R.L. Thomas, Rev. Sci. Instrum., Vol 71, 2000, p 2418 12. J. Rantala, D. Wu, and G. Busse, Amplitude Modulated Lock-in Vibrothermography for NDE of Polymers and Composites, Research in Nondestructive Evaluation, Vol 7, No. 4, 1995, p 215–228 13. I. Perez and W.R. Davis, Optimizing the Thermosonic Signal, Review of Progress in QNDE, Vol 23, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 2003

A.V. Bray, Corrosion Monitoring Using Microwave and Guided Wave Nondestructive Evaluation, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 529–532

Corrosion Monitoring Using Microwave and Guided Wave Nondestructive Evaluation Alan V. Bray, Systems and Materials Research Consultancy (SMRC)

Introduction TWO INTERESTING TECHNIQUES for finding hidden corrosion are emerging as viable in-service nondestructive evaluation (NDE) tools: microwave NDE and guided wave (GW) ultrasonic NDE methods. The in-service inspection scenario is complicated by the fact that corrosion is often hidden, and expensive disassembly can be required to complete an inspection. Microwave and GW NDE techniques have promise as methods that minimize disassembly in many common corrosion detection problems. Both methods are capable of detecting corrosion damage, cracks, and other defect types in inaccessible areas. The most common corrosion-detection application of microwave NDE is inspecting under a coating or appliqué for corrosion. This corrosion detection is based on the physics of the way in which microwaves interact with the materials of the coating and the sublayers of corrosion. Paint, appliqués, and corrosion byproducts are all dielectric materials, and each influences a microwave signal differently. In addition, the damage caused by corrosion can change the distance from the coated surface to the base metal, and this change in height is also detectable. Aircraft coatings are critical targets because corrosion under a coating can ultimately result in loss of airframe integrity. Non-corrosion-related applications include inspection of composites and detection of moisture in nonconducting structures. The NDE by GW differs from other ultrasonic methods in two distinct ways: GW evaluation is typically conducted at lower frequencies—often under 1 MHz, and GW signals travel long distances. Thus an area can be inspected from a significant distance. Some GW methods are aimed at detecting corrosion wastage from as far away as 20 m (65 ft) from a GW sensor array. Applications include inspection of insulated pipe runs while only removing a small portion of the insulation, and inspection of structures such as beams from a fixed set of GW sensors. Like many NDE techniques, microwave and guided wave ultrasonic methods can be valuable tools for detecting hidden corrosion when used properly and interpreted with care. Much of the current research in applying these NDE methods to modern inspection problems centers on developing systems that are insensitive to errors and false alarms, and this often means systems that are specific to a given inspection scenario. As an example, a guided wave inspection system for insulated piping would be quite different than one used for detecting the presence of wing icing. Similarly, microwave NDE systems for detecting corrosion under coatings are configured differently than for moisture detection within a nonconducting structure like a honeycomb panel or a sandwich panel. Many of the basic system tools remain unchanged, but sensor configurations, frequency ranges, and analysis techniques can differ dramatically from one application to the next.

A.V. Bray, Corrosion Monitoring Using Microwave and Guided Wave Nondestructive Evaluation, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 529–532 Corrosion Monitoring Using Microwave and Guided Wave Nondestructive Evaluation Alan V. Bray, Systems and Materials Research Consultancy (SMRC)

Microwave NDE Devices Microwave transmitters provide the incident energy for an inspection. The most commonly used transmitters for corrosion detection are actually transceivers; diodes in the transmitter receive the reflected signal. The incident and reflected signals are separated in the device, and the reflected signals are then analyzed for evidence of corrosion. The presence of corrosion attenuates the received signal, as do changes in the distance from the metal substrate caused by wastage. A key to microwave NDE device design is keeping the sensor at a fixed distance from the inspection surface in order to avoid misinterpreting changes in standoff distance from actual corrosion features. The mechanical design, coupled with electronic processing of the reflected signal, are crucial in minimizing false alarms. Open-ended waveguides are used often as microwave corrosion NDE sensors, although open- ended coaxial sensors (Ref 3) and strip-line antennae (Ref 4) have also successfully detected defects under coatings. Open ended rectangular waveguide sensors typically operate in the X- band (8 to 12.5 GHz) or K-band (12.5 to 40 GHz) frequency range. Corrosion detection sensitivity is dependent on frequency and standoff distance. Emerging microwave NDE tools for inspecting under coatings include small programmable logic devices that aid in distinguishing good and bad operating conditions (e.g., conditions that would lead to false alarms or missed detections). Microwave devices sense the dielectric change that results when corrosion products are present under the coating. The amount/degree of corrosion has an influence that can also be sensed by the device. A heavily corroded area will attenuate the signal output more than a lightly corroded area will. Device displays can include a range of corrosion severity based on the extent of change in the returned signal. Defining the amount or degree of corrosion at the threshold of detection is difficult since it is defined by the corrosion itself. Some groups have used standard exposure methods such as ASTM G 34 specimens to define the lower detection limits in terms of hours of exposure under a defined regimen of corrosion (Ref 5). Inspections with a microwave NDE device can be done with standard scanning systems, but handheld scanning is also possible. Any microwave scanning mechanism must be designed with altitude (standoff) distance and cant angle sensitivity in mind. Robust systems include logic to eliminate drastic changes that could be changes in altitude. As an example, the radius of curvature of the inspection surface can cause inaccuracies if the altitude control system has a wide enough footprint so that the distance to the sensor changes significantly with radius of curvature of the part. Principles of Operation. Microwave NDE techniques depend on the interaction of incident microwave energy with a material or structure. This interaction involves the effect of the material on the electric and magnetic fields generated by the microwave device. Changes in the conductivity, permittivity, and permeability of the inspected material relative to its surroundings are sensed in these field changes, as well as changes in substrate morphology. Figure 1 illustrates that the damage caused by corrosion of the substrate and the corrosion product itself both result in a change in the received signal. The bottom sketch portrays change caused by a change in the distance to the substrate, the top attenuation of the signal caused by the presence of corrosion products. Clearly, most detections are some combination of the two extremes. Corrosion products are typically dielectric materials, such as metal oxides—for example, iron and aluminum oxides. The dielectric nature of a material is described mathematically by its permittivity and loss factor, the real and imaginary components of the dielectric characteristic. Microwave NDE techniques for detecting corrosion under paints and appliqués exploit the contrast in the dielectric characteristics of corrosion versus that of the coating as well as the changes in substrate characteristics that indicate corrosion damage such as pitting and wastage.

Fig. 1 The same incident signal (I1) yields differing reflected signals (R1 and R2). R1 is the return from an undamaged area, the base-line measurement. On the top, the change from R1 to R2 is due to the dielectric nature of the corrosion product. On the bottom the change is due to a combination of the dielectric change and a change in the transmission length. In the microwave sensor, a three-port circulator or similar element is used to separate the incident and reflected signals. The drive signal for the microwave source is typically a continuous wave transmitter. These are lowpower devices, typically under 0.5 mW; however, shielding the operator from stray microwave radiation is considered a good safety practice. Corrosion can be detected as an attenuation of the received signal relative to that for a noncorroded state. Detecting Hidden Corrosion. Since microwaves are reflected by metals, they are not useful in inspecting under metal structures such as lap joints. Their primary use as a corrosion NDE tool is in inspecting under paints and appliqués applied to metallic substrates. Microwave NDE methods can be used to inspect deep into the material matrix of nonconducting structures such as fiberglass composites (e.g., Ref 6) for delaminations, voids, and the presence of moisture. When scanning over a painted surface, corrosion products under the paint cause a change in the electromagnetic energy reflected back to the microwave device. It is this change that is then used to indicate the presence of corrosion (Ref 7). Microwaves are very sensitive to the presence of moisture, and this feature can be both a hindrance and a boon in NDE. Surfaces that are wet cannot be inspected without drying, yet the ability to find trapped moisture can be a valuable asset in some applications. Another scanning issue is inspection in corners and tight spots. The mechanical design used to maintain standoff control is rarely small enough to get into these tight spaces, yet these areas are a high priority in NDE because of their potential for corrosion. Special heads can be fitted to the wave guide of the device to maintain standoff control in corners, but it is difficult to include all the possible corner configurations in a single NDE device kit. In operations where the inspection geometry is constant from one unit to the next, such as inspection of a single aircraft type, a special set of heads can be developed to cover the various corners and tight spots. In-service inspections can be done with a handheld device, and the power consumption is low enough for battery operation. Scanning speeds on a uniform surface such as an aircraft skin are surprisingly fast and are only limited by the electronic processing time. Scanning is typically done in a quick search mode that looks for any indication of degradation, and then the indicated area is inspected more carefully to ensure that it is not a false alarm and to get a rough determination of the severity and size of the corroded area. Typical Applications. Microwave corrosion NDE applications that concentrate on inspection of the metallic substrate under coatings include shipboard painted or laminated structures (Ref 8), painted and appliquéd

aircraft skins (Ref 5), coated bridge structures (Ref 9), or any other metal component under a coating that would otherwise be difficult to inspect without removing the coating. Microwave sensors used for corrosion NDE are most often single-point devices. Focused microwave sensors have been successfully used in other applications such as inspection of fiber-reinforced plastic (FRP) jacketed concrete structures (Ref 10). Arrays of microwave sensors for inspecting large areas such as aircraft skins have potential applications, but sensor control (standoff and cant angle) can become problematic. A spot microwave NDE sensor is a relatively simple device, and inspection rates are fast compared with those for most other NDE techniques. The reason for this is that there is very little integration time required to make a corrosion/no- corrosion decision. Scan speeds of 0.3 to 0.6 m/ s (1 to 2 ft/sec) have been reported, which is roughly the hand speed of moving a paint brush. A microwave NDE device with a 12 mm (0.5 in.) aperture can easily cover over 20 to 40 m2 (200 to 400 ft2) in an hour in the hands of a trained operator. Operator displays can be as simple as looking at a string of light-emitting diodes, hearing an audible alarm, or watching a meter. Even with modern digital logic built in, it is still important that an operator be able to discern the difference between changes in sensor altitude caused by rivets or seams, and changes due to corrosion. The low cost of microwave NDE sensors is attractive, but as with any NDE procedure, training is essential to ensure reliable corrosion detection.

References cited in this section 3. R. Zoughi, C. Huber, J. Easter, and K. Hayes, Microwave Detection of Surface Cracks in Metals Under Dielectric Coatings Using Open-Ended Coaxial Sensors: Preliminary Results, ASNT, Topics on Nondestructive Evaluation Series, Vol 2 (Jan 1998), p 53–60 4. J.A. Navarro and K. Chang, Inverted Stripline Antennas Integrated with Passive and Active Solid-State Devices, Trans. Microwave Theory Techniques, Vol 43 (No. 9), Sep 1995, p 2069–2065 5. A.V. Bray, H. Stretz, and S.C. Buckner, Coating Effects in Microwave NDE in the Detection of Corrosion in Aircraft, Paper 228-6200, presented in Proc. National Association of Corrosion Engineers Symposium 2000 (Orlando, FL), NACE International, March 2000 6. J. M. Liu, Microwave Scattering for the Characterization of a Disc-Shape Void in Dielectric Materials and Composites, Proc. of the 24th Annual Symposium on Quantitative Nondestructive Evaluation, July 1997, p 699–703 7. N. Qaddoumi, L. Handjojo, T. Bigelow, J. Easter, A. Bray, and R. Zoughi, Microwave Corrosion Detection Using Open Ended Rectangular Waveguide Sensors, Mater. Eval., Vol 58 (No. 2), February 2000, p 178–184 8. A.V. Bray, K.E. Jenne, N. Qaddoumi, J.K. Easter, T. Bigelow, and R. Zoughi, Microwave NDE Methods for Detecting Corrosion Under Paint, Proc. of the American Society of Naval Engineers Condition Based Maintenance Symposium, (Arlington, VA), June 1998. Available http:// www.navalengineers.org/Publications/ Pubs.html 9. R. Zoughi, Principles of and Applications of Microwave and Millimeter Wave NDE Methodologies: An Overview, Nondestructive Evaluation of Materials and Composites, Dec 1996, p 46–54 10. Y.J. Kim, F. Flavius, L. Jofre, and M.Q. Feng, Proc. of the 46th Annual Society for the Advancement of Material and Process Engineering Symposium, (Long Beach, CA), May 2001. Available http:// www.sampe.org/publicat.html#symposium


Guided Wave Ultrasonic Devices Guided wave ultrasonic NDE methods have a loose microwave analog—the wave guide. Guided waves are generated by first flooding a structure with acoustic energy. This energy is reflected from the boundaries of the structure with small changes in frequency, and these signals constructively combine to form wave packets that travel within structure much like a wave guide. Some wave packets, or modes, can travel long distances without significant attenuation once they are formed (Ref 11), and these are the richest modes for detecting corrosion wastage. Figure 2 illustrates the manner in which GW modes are formed in a structure. There are many variations on the basic elements of GW systems of NDE; these include a wide range of frequency selection, drive pulse shapes, sensor arrays, and electronic display systems.

Fig. 2 The formation of a guided wave within a structure is shown on the left. Superposition of the component waves create the wave packet within the carrier envelope. Guided waves are useful in inspection scenarios where an item is covered or encased. Examples include insulated piping, sandwiched structures such as lap joints, encapsulated pressure vessel walls, and applications such as wing icing where a guided wave is generated in the wing skin (Ref 12) and the presence of ice shows up as a change in the return signal. Guided wave NDE systematic “noise” sources include mode- coupling phenomena, array pattern dropouts or blind spots, and operator training and interpretation. Principles of Operation. Most GW ultrasonic sensors are piezoelectric transducers in either crystalline or piezocomposite format. Guided wave transducers are driven by high- voltage (on the order of 1000 V in some applications) tone burst electronics. Detection of corrosion damage is done in pulse-echo or throughtransmission modes, the latter requiring a separate receiving sensor or array. Common displays include amplitude versus time of arrival for a given frequency range. Each mode travels at a different speed, but in most inspections the frequency range is narrow enough that the time scale can be recalibrated in distance from the transducer. Guided wave ultrasonic NDE sensors can be classified into two major designs—those that are in contact with the inspection surface and those that use standoff sensors such as electromagnetic acoustic transducers (EMATs) or laser generated/ received sound (Ref 13). Contact sensors can be used in such applications as insulated piping systems, but a section of the insulation must be removed from the transmitter and receiver sites. Electromagnetic acoustic transducers and laser devices can insert and receive sound through porous coatings up to the limiting distance for the device.

In many guided wave applications, arrays of GW sensors are driven in phase or with programmed delays for activating sensors in the array. The advantages of an array include larger cross-sectional coverage, increased distances of inspection, and the ability to tune the signal using array delay times to increase the probability of detecting corroded or cracked areas. Detecting Hidden Corrosion. Pipe or pressure vessel wall thinning due to corrosion is a classic GW target. As the GW packets strike a thinned area, mode changes are induced that can be detected as changes in frequency content and time delay. Time delay changes result since each mode travels at a unique velocity. Mode diagrams of phase velocity versus f × d, the product of frequency (f) and the dimension (d) of the “waveguide.” In a pipe inspection, d is the pipe wall thickness. These mode diagrams in the frequency region of interest are an important tool in understanding inspection results. Tone burst systems are typically built on a personnel computer platform, and many of the analysis functions are programmed into the system. The smallest defects that are detectable with NDE systems by GW method depend on the application. In pipe inspection scenarios, the lower defect limit is usually characterized as 5 to 7% of the cross-sectional area of the pipe wall (Ref 14). Since the distance of the defect from the transducer is known, it is a simple matter to pinpoint the defect location with a tape measure. This makes it an excellent tool for quickly scanning large lengths of pipe, tubing, and hidden aircraft structures to identify and locate in-service corrosion problems that can then be examined in detail with more conventional ultrasonic methods. Typical Applications. Guided wave ultrasonic applications encompass a wide range of customized scenarios for corrosion and crack NDE. Monitoring existing structures with an array of guided wave leave-in-place (LIP) sensors is a relatively new technique (Ref 15). Small GW sensors are placed in the vicinity of a critical high stress region, and initial measurements record the state of the part at that point in time. As the part is used in service the “array” that was formed by the LIP sensors can be interrogated and compared to the initial state. This is a particularly pertinent technique for parts that have a history of failure under stress. Leave-in-place sensors for monitoring existing structures operate in the 200 KHz region. The sensors are relatively inexpensive, and they can be left in place for years. As frequencies drop to accommodate larger structures, the GW sensor size naturally increases. In applications requiring a lower frequency approach, the sensor array is commonly moved from site to site and operated in pulse- echo mode. Trend analysis of inspection results is possible, but care must be taken to record the original array positions if time/frequency scans are to be compared over time. The pipe inspection application is an example of a low frequency GW NDE application employing this approach (Ref 16).

References cited in this section 11. D.N. Alleyne and P. Cawley, Long Range Propagation of Lamb Waves in Chemical Plant Pipework, Mater. Eval., Vol 55 (No. 4), April 1997, p 504–508 12. D.D. Hongerholt, G. Willms, and J.L. Rose, “Summary of Results from an Ultrasonic In- Flight Wing Ice Detection System,” presented at American Society for Nondestructive Testing (ASNT) Quantitative Nondestructive Evaluation, Brunswick, ME, July 2001 13. Z. Guo, J.D. Achenbach, and S. Krishnaswamy, EMAT Generation and Laser Detection of Single Lamb Wave Modes, Proc. of the 24th Annual Symposium on Quantitative Nondestructive Evaluation, July 1997, p 201–208 14. P.J. Mudge, “Field Application of the Teletest Long-Range Ultrasonic Testing Technique,” Insight, Vol 43 (No 2), Feb 2001 15. J.L. Rose, and T. Hay, “New Technological Challenges in Ultrasonic Guided Wave NDE,” presented at Tri-Service Corrosion Conference (San Antonio) 14–18 Jan 2002 16. A.V. Bray, L.H. Carbaugh, R.K. Austin, J.L Rose, M. Avioli, and M. Quarry, Detecting Corrosion Under Insulation In Shipboard Piping II, Proc. of the U.S. Navy and Industry Corrosion Technology Information Exchange (Louisville, KY), July 1999


References 1. R. Zoughi and S. Ganchev, Microwave Nondestructive Evaluation, A State of the Art Review, NTIAC 95-01, Nondestructive Testing Information Analysis Center. Available http://www.ntiac.com/pubs/micro.html 2. J.L. Rose, Ultrasonic Waves in Solid Media, Cambridge University Press, 1999 3. R. Zoughi, C. Huber, J. Easter, and K. Hayes, Microwave Detection of Surface Cracks in Metals Under Dielectric Coatings Using Open-Ended Coaxial Sensors: Preliminary Results, ASNT, Topics on Nondestructive Evaluation Series, Vol 2 (Jan 1998), p 53–60 4. J.A. Navarro and K. Chang, Inverted Stripline Antennas Integrated with Passive and Active Solid-State Devices, Trans. Microwave Theory Techniques, Vol 43 (No. 9), Sep 1995, p 2069–2065 5. A.V. Bray, H. Stretz, and S.C. Buckner, Coating Effects in Microwave NDE in the Detection of Corrosion in Aircraft, Paper 228-6200, presented in Proc. National Association of Corrosion Engineers Symposium 2000 (Orlando, FL), NACE International, March 2000 6. J. M. Liu, Microwave Scattering for the Characterization of a Disc-Shape Void in Dielectric Materials and Composites, Proc. of the 24th Annual Symposium on Quantitative Nondestructive Evaluation, July 1997, p 699–703 7. N. Qaddoumi, L. Handjojo, T. Bigelow, J. Easter, A. Bray, and R. Zoughi, Microwave Corrosion Detection Using Open Ended Rectangular Waveguide Sensors, Mater. Eval., Vol 58 (No. 2), February 2000, p 178–184 8. A.V. Bray, K.E. Jenne, N. Qaddoumi, J.K. Easter, T. Bigelow, and R. Zoughi, Microwave NDE Methods for Detecting Corrosion Under Paint, Proc. of the American Society of Naval Engineers Condition Based Maintenance Symposium, (Arlington, VA), June 1998. Available http:// www.navalengineers.org/Publications/ Pubs.html 9. R. Zoughi, Principles of and Applications of Microwave and Millimeter Wave NDE Methodologies: An Overview, Nondestructive Evaluation of Materials and Composites, Dec 1996, p 46–54 10. Y.J. Kim, F. Flavius, L. Jofre, and M.Q. Feng, Proc. of the 46th Annual Society for the Advancement of Material and Process Engineering Symposium, (Long Beach, CA), May 2001. Available http:// www.sampe.org/publicat.html#symposium 11. D.N. Alleyne and P. Cawley, Long Range Propagation of Lamb Waves in Chemical Plant Pipework, Mater. Eval., Vol 55 (No. 4), April 1997, p 504–508 12. D.D. Hongerholt, G. Willms, and J.L. Rose, “Summary of Results from an Ultrasonic In- Flight Wing Ice Detection System,” presented at American Society for Nondestructive Testing (ASNT) Quantitative Nondestructive Evaluation, Brunswick, ME, July 2001

13. Z. Guo, J.D. Achenbach, and S. Krishnaswamy, EMAT Generation and Laser Detection of Single Lamb Wave Modes, Proc. of the 24th Annual Symposium on Quantitative Nondestructive Evaluation, July 1997, p 201–208 14. P.J. Mudge, “Field Application of the Teletest Long-Range Ultrasonic Testing Technique,” Insight, Vol 43 (No 2), Feb 2001 15. J.L. Rose, and T. Hay, “New Technological Challenges in Ultrasonic Guided Wave NDE,” presented at Tri-Service Corrosion Conference (San Antonio) 14–18 Jan 2002 16. A.V. Bray, L.H. Carbaugh, R.K. Austin, J.L Rose, M. Avioli, and M. Quarry, Detecting Corrosion Under Insulation In Shipboard Piping II, Proc. of the U.S. Navy and Industry Corrosion Technology Information Exchange (Louisville, KY), July 1999



J.L. Rose, Ultrasonic Waves in Solid Media, Cambridge University Press, 1999 R. Zoughi and S. Ganchev, Microwave Nondestructive Evaluation, A State of the Art Review, NTIAC 95-01, Nondestructive Testing Information Analysis Center. Available http:/ /www.ntiac.com/

S.W. Dean, Corrosion Monitoring for Industrial Processes, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 533–541

Corrosion Monitoring for Industrial Processes Sheldon W. Dean, Dean Corrosion Technology, Inc.

Introduction IN-SERVICE MONITORING of industrial manufacturing operations is the type of service- corrosion testing that presents the greatest challenge and for which there is a great need. In such operations, the expense of corrosion problems can be huge and the risks devastating. These expenses are summarized in the article “Direct Costs of Corrosion in the United States” in this Volume. The detailed report to the United States Congress is available (Ref 1). Corrosion monitoring has become an important aspect of the design and operation of modern industrial plants because it is one of several techniques that can be used to keep track of the degradation processes that may be occurring in an industrial operation. The most widely used application of in-service monitoring is to identify the occurrence of corrosive conditions that may be caused by unknown circumstances, such as a new catalyst that promotes corrosive chemicals formation, or a feed stock that may contain corrosive

impurities from time to time. In such circumstances, in-line monitoring can be an important tool to maintaining safe operation. Because of the high cost of installing in-line monitoring systems, and the need for regular maintenance, these systems are seldom used in well-behaved operations. In these cases, periodic inspections suffice to meet both the regulatory and management needs for a safe operation. However, in situations such as oil and gas production, where substantial variations in process fluids may occur in a relatively short time, the value of on-line monitoring is higher and can be essential to long- term safety and reliability. This article, therefore, focuses on methods of monitoring corrosion in industrial plants. The term monitoring, as used in this context, includes any technique for evaluating the progress or rate of corrosion.

Reference cited in this section 1. G.H. Koch, M.P.H. Brongers, N.G. Thompson, Y.P. Virmani, and J.H. Payer, “Corrosion Cost and Preventive Strategies in the United States,” FHWA-RD-01-156, Federal Highway Administration, U.S. Department of Transportation, Washington D.C., March 2002

S.W. Dean, Corrosion Monitoring for Industrial Processes, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 533–541 Corrosion Monitoring for Industrial Processes Sheldon W. Dean, Dean Corrosion Technology, Inc.

Selecting a Corrosion-Monitoring Method A large variety of techniques are available for corrosion monitoring in plant corrosion tests, and much has been written on the subject (Ref 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12). The most widely used and simplest method of corrosion monitoring involves the exposure and evaluation of the actual test coupons (specimens). ASTM G 4 was developed to provide guidance for this type of testing (Ref 13). Additional detailed information on procedures from practical experience is given in Ref 10. Extensive overviews of some electrochemical methods of corrosion monitoring in industrial plants are given in Ref 7 and 8. ASTM G 96 (Ref 14) provides detailed information on the use of electrical and electrochemical monitoring methods. However, in view of the growing number of methods available, the selection of corrosion-monitoring methods may be somewhat arbitrary. In the selection of a corrosion-monitoring method, a variety of factors should be considered. First, the purpose of the test should be understood by everyone concerned with the corrosion-monitoring program. The cost and applicability of the methods under consideration should be known, and it is important to consider the reliability of the method selected. In many cases, it is desirable to include more than one method in order to provide more confidence in the information generated. Another key question is what type of access to the process streams and equipment is available. If access is available, methods that involve probes or coupons become more feasible. Otherwise, nondestructive methods may be required. An important factor in the selection of monitoring methods is the response time required to obtain the desired information from the method. Coupon methods and techniques that require plant shutdown for installation and retrieval of specimens tend to be relatively slow in generating information. On the other hand, equipment that measures instantaneous corrosion rates can provide rapid results, but often with a reduction in accuracy. A most important consideration is safety. In an operating plant, equipment failure can lead to a leak, which can result in loss of product, a hazard potential, and possible shutdown of the plant. It is important for the monitoring apparatus to be designed to minimize the possibility of such an incident and to have a plan in place to safely deal with any leak that might develop.


2. C.P. Dillon, A.S. Krisher, and H. Wissenberg, Plant Corrosion Tests, Handbook on Corrosion Testing and Evaluation, W.H. Ailor, Ed., John Wiley & Sons, 1971, p 599–615 3. A.S. Couper, Methods in Oil-Refining Industries, Handbook on Corrosion Testing and Evaluation, W.H. Ailor, Ed., John Wiley & Sons, 1971, p 617–624 4. “Laboratory Corrosion Testing of Metals for the Process Industries,” NACE TM-01-69, National Association of Corrosion Engineers, 1976 5. J.N. Wanklyn, Corrosion Monitoring in the Oil and Petrochemical and Process Industries, J. Wanklyn, Ed., Oyez Scientific and Technical Service Ltd., 1982 6. Corrosion Monitoring in Industrial Plants Using Nondestructive Testing and Electrochemical Methods, STP 908, G.C. Moran and P. Labine, Ed., American Society for Testing and Materials, 1986 7. S.W. Dean, Overview of Corrosion Monitoring in Modern Industrial Plants, Corrosion Monitoring in Industrial Plants Using Nondestructive Testing and Electrochemical Methods, STP 908, G.C. Moran and P. Labine, Ed., American Society for Testing and Materials, 1986, p 197–220 8. S.W. Dean, Electrochemical Methods for Corrosion Testing in the Process Industries, Electrochemical Corrosion Testing with Special Consideration of Practical Applications, E. Heintz, J.C. Rowlands, and F. Mansfeld, Ed., Proceedings of International Workshop, (Ferrara, Italy), DECHEMA, 1986, p 1–15 9. Process Industries Corrosion—Theory and Practice, National Association of Corrosion Engineers, 1986 10. B.J. Moniz, Field Coupon Corrosion Testing, Process Industries Corrosion—Theory and Practice, National Association of Corrosion Engineers, 1986 11. S.J. Pikul, T.S. Lee, and D.B. Anderson, “Field Corrosion Tests for Materials Evaluation in Chemical Process Industries,” Paper 119, presented at Corrosion/86, National Association of Corrosion Engineers, 1986 12. A. Perkins, Industrial Applications, Corrosion Tests and Standards, Manual 20, ASTM International, 1995, p 143–148 13. “Standard Method for Conducting Corrosion Coupon Tests in Plant Equipment,” G 4, Annual Book of ASTM Standards, Vol 03.02, ASTM International 14. “Standard Guide for On-Line Monitoring of Corrosion in Plant Equipment (Electrical and Electrochemical Methods),” G 96, Annual Book of ASTM Standards, Vol 03.02, ASTM International


Direct Testing of Coupons

Although they are not intended to replace laboratory tests, plant coupon tests usually are designed to monitor the damage rate occurring on existing equipment, to evaluate alternative materials of construction, and sometimes to determine the effects of process conditions that cannot be reproduced in the laboratory. Advantages of Coupon Testing (Ref 10). Plant coupon testing provides several specific advantages over laboratory coupon testing. A large number of materials can be exposed simultaneously and can be ranked in actual process streams with actual process conditions. Testing can be used to monitor the corrosivity of process streams. The coupons can be designed for specific forms of corrosion, but the exposure time is usually determined by the plant operation. Large Number of Coupons. Because many coupons can be exposed simultaneously, they can be tested in duplicate or triplicate (to measure scatter), and they can be fabricated to simulate such conditions as welding, residual stresses, or crevices. These variations can provide the engineer with increased confidence in selecting materials for new equipment, maintenance, or repair. Actual process streams reveal the synergistic effects of combinations of chemicals or contaminants. In addition, the possibility is remote that the corrosion of specific coupons contaminates the process and affects the corrosion resistance of other coupons. However, one potential source of error that must be checked is contamination of the process stream by the corrosion of the existing equipment. For example, in a hypothetical case, the equipment is fabricated from a nickel-base alloy, and the process is a reducing acid. Contamination of the process by nickel ions may result in an apparent improvement in the corrosion resistance of titanium test coupons. If the existing equipment is then replaced with titanium, based on these misleading test results, it may corrode rapidly without the beneficial effect of the nickel ions. Monitoring of Inhibitor Programs. Coupons often are used to monitor inhibitor programs in, for example, water treatment or refinery overhead streams. With retractable coupon holders, the coupons can be extracted from the process without having to shut down in order to determine the corrosion rate. Long Exposure Times. Some forms of localized corrosion, such as crevice corrosion, pitting, and stresscorrosion cracking (SCC), require time to initiate. To increase confidence in the test results, coupons can be exposed for as long as necessary to ensure that the initiation time has been exceeded. Coupon design can be selected to test for specific forms of corrosion. Coupons can be designed to detect such phenomena as crevice corrosion, pitting, and dealloying corrosion. For example, some pulp mill bleach plant washer drums are protected electrochemically to mitigate crevice corrosion. Specifically designed crevicecorrosion test coupons are used to monitor the effectiveness of the electrochemical protection program. These coupons are removed periodically from the equipment and examined for evidence of crevice corrosion. Disadvantages of Coupon Testing (Ref 10). Coupon testing has four main limitations: • • •

•

Coupon testing cannot be used to detect rapid changes in the corrosivity of a process. Localized corrosion cannot be guaranteed to initiate before the coupons are removed— even with extended test durations. Corrosion rate calculated from coupons may not reflect the corrosion of the plant equipment because of factors such as multiphase flow, where the aqueous phase is much more corrosive than the organic or vapor phase, or turbulence from mixers, elbows, valves, pumps, and other items that accelerate the corrosion in a specific location in the equipment removed from where the coupons were exposed. Coupons may be misleading in situations where the corrosion rate varies significantly over time because of unrealized process factors.

Contamination of the process fluid is an issue that must be addressed in the food, medical, and electronic equipment manufacturing industries, if coupons are introduced. Rapid Changes in Corrosivity. The corrosion rate calculated from a coupon is an average over a specific period of time. Therefore, field coupon testing cannot detect process upsets as they occur. For real-time monitoring, electrochemical methods, such as the polarization-resistance technique, may be more appropriate (see the section “Polarization-Resistance Measurement” in this article). Corrosion Rates of Coupons Compared with That of Equipment. The corrosion rate of plant equipment seldom equals that calculated on a matching test coupon because it is almost impossible to duplicate the equipment exposure with a coupon exposure. However, coupons are very helpful in many situations because they provide a clear view of the corrosion damage with information on the development of deposits, films, and localized

attack. In addition, coupon testing is probably the least complicated type of in-service testing and, consequently, requires the least operator skill. Corrosion Forms Not Detected. The principal limitation of coupon testing is the simulation of erosion-corrosion and heat-transfer effects. Careful placement of the coupons in the process equipment can slightly offset these weaknesses. Erosion-corrosion is related to process turbulence, and process turbulence is often a function of equipment design. Because coupons tend to shield one another from the effects of process turbulence, field coupon testing is not reliable as a method of simulating erosion-corrosion. For heat-transfer effects, specially designed coupons are required that simulate effects such as those found in heating elements or condenser tubes. Coupons range in design from thermowell-shaped devices to sample tubes in a test heat exchanger. Thermowell-shaped devices are heated or cooled on the inside and project into the process stream (Ref 13). Heat-transfer tests can also be conducted in the laboratory. In this environment, the coupon forms part of the wall of the test vessel and can therefore be heated or cooled from one side. Because of the cost involved, heat-transfer coupon tests are usually carried out on only one (or perhaps two) alloys that have been selected from a larger group. Coupon Options (Ref 10). The design of the coupon is an important part of any plant corrosion-testing program. Proper selection of the coupon shape, surface finish, metallurgical condition, and geometry allows evaluation of specific forms of corrosion. Uniform Corrosion Coupons. The most common coupon shape for the evaluation of uniform corrosion is rectangular. Circular shapes are also used. Rectangular coupons are the most common because most alloys are available in plate or sheet form. Other shapes are used when there are restrictions on available product forms or when a specific material condition is required. Coupon identification must be legible and permanent. The simplest, and preferred, method of identification is stencil stamping. Coupon finish represents a significant contribution to the overall cost. The least expensive finish that is consistent with the test requirements should be selected. For example, an inexpensive surface finish is acceptable where carbon steel coupons are used routinely to monitor additions of inhibitor in water treatment programs. This may be achieved by punching or shearing, followed by glass bead blasting. On the other hand, when it is necessary to rank alloys in a process environment, the coupons must be finished with ground or machined parallel edges and sanded faces. A wide variety of coupon finishes are available, including: • • • • • • •

Mill finish Glass bead blasted Sandblasted Steel shotblasted Abrasive cloth or paper sanded Machined Electrolytically polished

Ideally, the surface finish of the coupon should match the finish of the equipment. This match is very difficult to achieve for several reasons. For example, there is the difference in surface finish between different mill product forms and even between different heats of the same metal or alloy. In addition, there can be wide variation among mill scale and surface deposits from processing operations. The principal requirement of testing is to assess the corrosion resistance of the test alloy in the condition that would be used in actual equipment. Abrasive cloth or paper sanding is the most common practice of surface preparation. Sanding removes such surface deposits as mill scale and such defects as scratches or pits. A 120-grit finish, which is standard, can be produced easily without specialized equipment. Metallurgical changes that are heat induced, such as the sensitization of stainless steels or nickel-base alloys, can be prevented by keeping the coupon cool through wet sanding with progressively finer abrasives. Clean polishing belts should be used to avoid contamination of the coupon surface. For example, a belt that has been used to polish a brass coupon should not be used to polish an aluminum coupon. The corrosion resistance of some alloys can be enhanced by a special surface condition, for example, the oxide films that are formed on titanium or zirconium. For these alloys, conditioning of the coupons would take place

after mechanical surface preparation. However, some films that form during mill processing or chemical exposure actually reduce corrosion resistance. Certain forms of iron sulfide on carbon steel are examples of these types of films. The coupons must be specially treated in such cases. Coupons that are cut by punching or shearing have cold-worked edges. The effects of cold work extend back from the cut edge a distance equal to the material thickness. These affected areas can be removed by grinding or machining. Cold-worked edges may affect the corrosion rate in some cases, and the residual stresses it induces may cause SCC in some materials. Extensive edge preparation can be a major contributor to the cost of the coupon. Galvanic-Corrosion Coupons. Pairs of test coupons are coupled electrically to study the effects of galvanic corrosion. The relative areas exposed usually vary from 1:1 to 10:1 or greater. The area ratios should be reversed for complete assessment, that is, 10:1 and 1:10. This reversal may not be necessary when a specific effect is being studied, for example, simulation of the galvanic corrosion of dissimilar-metal fasteners in column trays. A major concern with electrically coupled coupons is maintaining electrical continuity for the entire exposure period. Corrosion product films can wedge mechanically joined coupons apart, thereby eliminating the electrical contact and any galvanic corrosion effect. ASTM G 71 (Ref 15) provides valuable information on galvanic corrosion testing in both field and laboratory environments. With metals that can become embrittled by hydrogen absorption—for example, titanium, zirconium, tantalum, and hardenable steels—the cathodic (protected) member of the galvanic couple may be subject to the greater damage. However, the typical mass-loss measurements would not reveal such damage. Crevice-Corrosion Coupons. Equipment crevices, such as weld backing rings, tube-tubesheet joints, or flanged connections, are common sites for localized corrosion in the process environment. Many metals perform differently in crevices as opposed to unshielded areas. Behavior is dependent on several factors. These factors include how strongly oxygen reduction (cathode depolarization) controls the cathode reaction or how the crevice alters the bulk process chemistry by lowering the pH or concentrating aggressive species. The most imaginative form of coupon-corrosion testing is the simulation of crevice corrosion (Ref 16). The various techniques that can be used for crevice-corrosion testing include rubber bands, spot-welded lap joints, and wire wrapped around threaded bolts. Each crevice test creates a particular crevice geometry between specific materials and has a particular anode/cathode area ratio. Thus, no crevice-corrosion test is universally applicable (see the article “Evaluating Crevice Corrosion” in this Volume). The two most widely used crevice geometries in field coupon testing employ insulating spacers to separate and electrically insulate the coupons. Spacers are usually either flat washers or multiple-crevice washers. Either type of spacer can be made of materials ranging from hard ceramics to soft thermoplastic resins (Ref 17). ASTM G 78 (Ref 17) and G 48 (Ref 18) provide valuable information on crevice corrosion specimen design and should be consulted before attempting a crevice-corrosion test. Stress-Corrosion Coupons. Typical sources of sustained tensile stress that cause SCC of equipment in service are the residual stresses resulting from forming and welding operations and the assembly stresses associated with interference-fitted parts, especially in the case of tapered, threaded connections. Therefore, the most suitable coupons for plant tests are the self-stressed bending and residual-stress specimens described in the article “Evaluating Stress-Corrosion Cracking” in this Volume. Convenient coupons are the cup impression, Ubend (Ref 19), C-ring (Ref 20), tuning fork, and welded panel (Ref 21). The method of stressing for all of these coupons results in decreasing load as cracks form and begin to propagate. Therefore, complete fracture is seldom observed, and careful examination is required to detect cracking. Welded Coupons. Because welding is a principal method of fabricating equipment, the use of welded coupons is often desirable. Aside from the effects of residual stresses, the primary concern is the behavior of the weld bead and the heat-affected zone (HAZ). In some alloys, the HAZ becomes sensitized to severe intergranular (sometimes called knife-line) attack, and in certain other alloys, the HAZ is anodic to the parent metal. When possible, it is more realistic to remove welded coupons from production-sized weldments than to weld the small coupons. Because the thermal conditions in both the weld metal and the HAZ are a function of the number of weld passes, the metal thickness, the weld position, and the welding technique, it is not usually considered good practice to use welded coupons to assess the possibility of sensitization from welding. It is generally better to carry out a sensitizing heat treatment on an unwelded coupon before it is tested and then look for evidence of intergranular corrosion or cracking.

Sensitized Metal. Sensitization is a metallurgical change that occurs when certain austenitic stainless steels, ferritic stainless steels, nickel- base alloys, and other alloys are heated under specific conditions. This results in the precipitation of carbides or other intermetallic phases at grain boundaries, which reduces corrosion resistance. Any heat-inducing process (for example, stress relief of weldments) may cause sensitization, which is time and temperature dependent. There is a specific temperature range over which each particular alloy sensitizes rapidly. Sensitization leads to intergranular corrosion in specific environments. Welding is the most common cause of sensitization. However, welded coupons may not exhibit sensitization because they may be given insufficient weld passes (compared with actual process equipment). As a result, they spend insufficient time in the sensitizing temperature range, and susceptibility to intergranular corrosion may not be detected. An appropriate sensitizing heat treatment guarantees that any corrosion susceptibility induced by welding or heat treatment is detected. The optimal temperature and time ranges for sensitization vary for different alloys. For example, 30 min at 650 °C (1200 °F) is usually sufficient to sensitize American Iron and Steel Institute (AISI) type 316 stainless steel. Some of the corrosion-resistant aluminum- magnesium (5000-series) alloys containing 3 to 6% Mg are also subject to sensitization when heated at temperatures in the range of 65 to 175 °C (150 to 350 °F). Reference 22 describes a test method for susceptibility to this phenomenon. The effective time and temperature conditions depend on the alloy content and metallurgical condition (Ref 23). The advantage of using sensitized coupons is that, if corrosion occurs, they indicate a potential problem. The user is thus obliged to consider how the equipment will be welded and heat treated and to determine whether high-temperature excursions will occur during operation. Test Rack Design. It may be possible to expose corrosion coupons in stagnant or slowly flowing conditions by simply hanging them on an insulated wire or plastic cord, but this procedure is generally inadequate. Instead, specimen holders and test racks are usually used to support and insulate the coupons. These racks must hold coupons firmly in place to prevent mechanical damage and metal loss from causes other than corrosion. They must also electrically isolate the coupons from contact with each other and from the vessel itself to prevent unintentional electrochemical interactions. The basic types of racks are discussed briefly as follows. They are described in detail in Ref 2, 5, 10, and 12. Spool (Birdcage) Rack. A typical spool rack is useful in open vessels, such as reactors and tanks, for which access to the spool is readily achieved (Fig. 1). It has the advantage that a relatively large number of coupons can be accommodated.

Fig. 1 Illustration of typical spool-type coupon rack. Source: Ref 13 The insert rack is convenient for use in pipes or other units that have flanged connections that allow easy access to the system. For larger-diameter pipes or nozzles, a dutchman-type rack can be used. Such a rack consists of a suitable spool-type piece with the coupons mounted crosswise on a bar (Fig. 2). In both cases, the equipment must be shut down during installation and removal of the racks, or a by-pass line added to the system that allows the process fluid to be routed around the specimens while they are being inserted or removed. It is very important to have good-quality shut-off valves and drain and flush provisions if a bypass is to be used.

Fig. 2 Illustration of typical dutchman coupon rack. Source: Ref 13 A slip-in rack is designed to be inserted into and removed from equipment that is in operation. A retractable coupon holder makes this type of rack especially useful (Fig. 3). The slip-in rack requires a full-port gate valve and a suitable- sized nozzle to serve as a retraction chamber. A rod-shaped coupon holder is contained in the retraction chamber, which is flanged to the gate valve. The other end of the retraction chamber contains a packing gland through which the coupon holder passes. Coupons are mounted on the rod in the extended position and are then drawn into the retraction chamber. The chamber is bolted to the gate valve, which is then opened to allow the coupons to be slid into the process stream. The sequence is reversed to remove the test coupons. It is essential that the rod or handle be equipped with a restraining chain or other device to prevent the blowout of the specimen holder when retracting the specimens when the system is under pressure. Drain and flush ports are desirable additions to deal with the process fluid that is present in the retraction chamber before the specimens are removed. These ports are also useful to eliminate intrusion of air into the process fluid during insertion of the coupons.

Fig. 3 Illustration of retractable coupon holder. Source: Ref 10 Coupon Cleaning and Evaluation. The test coupons should be cleaned as soon as possible after removal from the test. The procedures for cleaning and weighing, which depend on the test material and the extent of corrosion, are described in Ref 4, 10, and 24. These methods are overviewed in the article “Statistical Interpretation of Corrosion Test Results” in this Volume. Examination of the coupons after cleaning and weighing reveals the forms of corrosion that may be expected in equipment made of the coupon material. Because the coupons used for calculating corrosion rates are usually small, they should be examined very carefully for signs of localized corrosion effects that could invalidate the mass- loss data. Coupons are examined with the unaided eye and then at increasing magnifications, up to 30 to 50×, with a binocular microscope. The scanning electron microscope is an extremely useful tool for detecting superficial localized effects. In some cases, coupons must be bent and/or sectioned and metallographically examined to reveal certain types of corrosion damage. There are special localized corrosion effects that could not only jeopardize determination

of realistic corrosion rates but also signal other serious types of behavior. These effects include edge attack, crevice corrosion, stress corrosion near stenciled identification numbers, pitting, selective corrosion (such as dealloying), knife-line attack, blistering, intergranular corrosion, embrittlement, and erosion-corrosion. In addition to the various corrosion effects, there may be inadvertent mechanical or galvanic corrosion damage. Techniques for evaluating specific forms of corrosion are described in various articles in this Section.

References cited in this section 2. C.P. Dillon, A.S. Krisher, and H. Wissenberg, Plant Corrosion Tests, Handbook on Corrosion Testing and Evaluation, W.H. Ailor, Ed., John Wiley & Sons, 1971, p 599–615 4. “Laboratory Corrosion Testing of Metals for the Process Industries,” NACE TM-01-69, National Association of Corrosion Engineers, 1976 5. J.N. Wanklyn, Corrosion Monitoring in the Oil and Petrochemical and Process Industries, J. Wanklyn, Ed., Oyez Scientific and Technical Service Ltd., 1982 10. B.J. Moniz, Field Coupon Corrosion Testing, Process Industries Corrosion—Theory and Practice, National Association of Corrosion Engineers, 1986 12. A. Perkins, Industrial Applications, Corrosion Tests and Standards, Manual 20, ASTM International, 1995, p 143–148 13. “Standard Method for Conducting Corrosion Coupon Tests in Plant Equipment,” G 4, Annual Book of ASTM Standards, Vol 03.02, ASTM International 15. “Standard Guide for Conducting and Evaluating Galvanic Corrosion Tests in Electrolytes,” G 71, Annual Book of ASTM Standards, Vol 03.02, ASTM International 16. W.D. France, Jr., Crevice Corrosion of Metals, Localized Corrosion—Cause of Metal Failure, STP 519, M. Henthorne, Ed., American Society for Testing and Materials, 1972, p 164–200 17. “Standard Guide for Crevice Corrosion Testing of Iron-Base and Nickel-Base Stainless Alloys in Seawater and Other Chloride- Containing Aqueous Environments,” G 78, Annual Book of ASTM Standards, Vol 03.02, ASTM International 18. “Standard Test Methods for Pitting and Crevice Corrosion Resistance of Stainless Steels and Related Alloys by Use of Ferric Chloride Solution,” G 48, Annual Book of ASTM Standards, Vol 03.02, ASTM International 19. “Standard Practice for Making and Using U- Bend Stress-Corrosion Test Specimens,” G 30, Annual Book of ASTM Standards, Vol 03.02, ASTM International 20. “Standard Practice for Making and Using C- Ring Stress-Corrosion Test Specimens,” G 38, Annual Book of ASTM Standards, Vol 0302, ASTM International 21. “Standard Practice for Preparation of Stress- Corrosion Test Specimens for Weldments,” G 58, Annual Book of ASTM Standards, Vol 03.02, ASTM International 22. “Standard Practice for Determining the Susceptibility to Intergranular Corrosion of 5xxx Series Aluminum Alloys by Weight Loss After Exposure to Nitric Acid (NAWLT Test),” G 67, Annual Book of ASTM Standards, American Society for Testing and Materials

23. E.H. Dix, Jr., W.A. Anderson, and M.B. Shumaker, Influence of Service Temperature on the Resistance of Wrought Aluminum-Magnesium Alloys to Corrosion, Corrosion, Vol 15 (No. 2), 1959, p 55t–62t 24. “Standard Practice for Preparing, Cleaning, and Evaluating Corrosion Test Specimens,” G 1, Annual Book of ASTM Standards, Vol 03.02, ASTM International


Electrical-Resistance Probes Electrical-resistance probes (Fig. 4) are specially designed corrosion coupons; see ASTM G 96, method A (Ref 14). Their corrosion rate is calculated from measurement of electrical resistance rather than mass loss. These measurements are made by installing a wire or other device fabricated from the material in question in such a way that its electrical resistance can be conveniently measured. Corrosion reduces the cross section of the exposed element; therefore, its electrical resistance increases with exposure time, if corrosion is taking place. A temperature- compensating element should be incorporated in such a probe because the resistance of the probe is also influenced by the temperature.

Fig. 4 Electrical-resistance probe. Source: Ref 7 When installing the probe, if galvanic corrosion is a concern, the probe should be tested for electrical isolation with adjacent metallic materials. Electrical-resistance probes measure the remaining average metal thickness. To obtain the corrosion rate, a series of measurements are made over a period of time, and the results are plotted as a function of exposure time. The corrosion rate can be determined from the slope of the resulting plot. There are several advantages to this approach. Because probes are relatively small, they can be installed easily. To determine the metal remaining, the probe can be wired directly to a control room location or to a portable resistance bridge at the probe location. For systems that are wired directly to control rooms, a computer system can be used to obtain the data and to reduce the results to corrosion rate values. Commercial electricalresistance equipment is available. Use of electrical-resistance probes does not depend on the environment; there are two reasons for this. First, only the remaining metal is measured, and second, the conductivity of the corrosive environment is usually inconsequential. This is not true in, for example, molten salts or liquid metals;

the high conductivity of the environment in such systems precludes the use of electrical-resistance probes. Thus, no problems exist in measuring corrosion that occurs in liquid or vapor streams. Electrical-resistance probes are used in various applications, such as monitoring the corrosivity of cooling water systems (Ref 14). However, there are some disadvantages to this approach. Vessel and piping walls must be penetrated in order to install the probes; such penetration results in the potential for leaks. It is expensive to direct wire the probe to a control room location, and such work must be carried out with care to avoid spurious signals and errors. On the other hand, it is time consuming and sometimes impossible to take measurements at the probe site with a portable bridge. The temperature-compensation device reacts slowly, and it can be a source of error if the temperature varies when the measurement is taken. Corrosion- rate measurements obtained in short periods of time can be inaccurate, because the method measures only the remaining metal, not the rate of attack; thus, the errors involved in estimating small differences between large numbers make this method inaccurate. This method provides no practical information on localized attack. It is also very important to select the correct type of probe for the application. Cylindrical probes or wires may not detect corrosion of a pipe wall from a thin film of liquid on the pipe wall resulting from annular two-phase flow. A flush-mounted probe must be used in this case. Another major problem is finding leak-tight insulators that have sufficient corrosion resistance to the process fluids encountered. In some cases, it has been necessary to weld a nozzle with a flange over the external connection so that the area could be closed tightly if a leak developed during operation.

References cited in this section 7. S.W. Dean, Overview of Corrosion Monitoring in Modern Industrial Plants, Corrosion Monitoring in Industrial Plants Using Nondestructive Testing and Electrochemical Methods, STP 908, G.C. Moran and P. Labine, Ed., American Society for Testing and Materials, 1986, p 197–220 14. “Standard Guide for On-Line Monitoring of Corrosion in Plant Equipment (Electrical and Electrochemical Methods),” G 96, Annual Book of ASTM Standards, Vol 03.02, ASTM International


Ultrasonic Thickness Measurements Ultrasonic thickness measurements can be used to monitor corrosion rates in situ. Ultrasonic thickness measurements involve placement of a transducer against the exterior of the vessel in question. The transducer produces an ultrasonic signal. This signal passes through the vessel wall, bounces off the interior surface, and returns to the transducer. The thickness is calculated by using the time that elapses between emission of the signal and its subsequent reception, along with the velocity of sound in the material. To obtain a corrosion rate, a series of measurements must be made over a time interval, and the metal loss per unit time must be determined. One advantage of this method is that it is not necessary to penetrate the vessel to make the measurements. In addition, the results are obtained in terms of thickness. Small, handheld probes and reading devices are available for making these measurements and are relatively easy to use. There are also several disadvantages. Ideally, a bare metal surface free of paint, thermal insulation, and corrosion products must be exposed. A coupling agent, such as grease, petroleum jelly, or oil, must be used so that the signal can pass from the probe into the metal and return. Problems may arise when vessel walls are at

high or low temperatures. Serious problems may exist in equipment that has a metallurgically bonded internal lining, because it is not obvious from which surface the returning signal originates. Despite these drawbacks, the ultrasonic thickness approach is widely practiced where it is necessary to evaluate vessel life and suitability for further service. It must be kept in mind, however, that depending on the type of transducer used, the ultrasonic thickness method can overestimate metal thickness when the remaining thickness is under approximately 1.3 mm (0.05 in.).


Ultrasonic Thickness Measurements Ultrasonic thickness measurements can be used to monitor corrosion rates in situ. Ultrasonic thickness measurements involve placement of a transducer against the exterior of the vessel in question. The transducer produces an ultrasonic signal. This signal passes through the vessel wall, bounces off the interior surface, and returns to the transducer. The thickness is calculated by using the time that elapses between emission of the signal and its subsequent reception, along with the velocity of sound in the material. To obtain a corrosion rate, a series of measurements must be made over a time interval, and the metal loss per unit time must be determined. One advantage of this method is that it is not necessary to penetrate the vessel to make the measurements. In addition, the results are obtained in terms of thickness. Small, handheld probes and reading devices are available for making these measurements and are relatively easy to use. There are also several disadvantages. Ideally, a bare metal surface free of paint, thermal insulation, and corrosion products must be exposed. A coupling agent, such as grease, petroleum jelly, or oil, must be used so that the signal can pass from the probe into the metal and return. Problems may arise when vessel walls are at high or low temperatures. Serious problems may exist in equipment that has a metallurgically bonded internal lining, because it is not obvious from which surface the returning signal originates. Despite these drawbacks, the ultrasonic thickness approach is widely practiced where it is necessary to evaluate vessel life and suitability for further service. It must be kept in mind, however, that depending on the type of transducer used, the ultrasonic thickness method can overestimate metal thickness when the remaining thickness is under approximately 1.3 mm (0.05 in.).


Measurement of Corrosion Potentials The use of corrosion-potential measurements for in-service corrosion monitoring is not as widespread as the use of polarization resistance. However, this approach can be valuable in some cases, particularly where an alloy could show both active and passive corrosion behavior in a given process stream. For example, stainless steels

can provide excellent service, as long as they remain passive. However, if an upset occurs that would introduce either chlorides or reducing agents into the process stream, stainless alloys may become active and may exhibit excessive corrosion rates. Corrosion-potential measurements would indicate the development of active corrosion, and they may be coupled with polarization-resistance measurements as additional confirmation of high corrosion rates (Ref 8). The success of corrosion-potential measurements depends on the long-term stable performance of a standard reference electrode. Such electrodes have been developed for continuous pH monitoring of process streams, and their application for measuring corrosion potentials is straightforward. However, the conditions of temperature, pressure, electrolyte composition, pH, and other possible variables can limit the applications of these electrodes for corrosion-monitoring service.

Reference cited in this section 8. S.W. Dean, Electrochemical Methods for Corrosion Testing in the Process Industries, Electrochemical Corrosion Testing with Special Consideration of Practical Applications, E. Heintz, J.C. Rowlands, and F. Mansfeld, Ed., Proceedings of International Workshop, (Ferrara, Italy), DECHEMA, 1986, p 1–15


Measurement of Corrosion Potentials The use of corrosion-potential measurements for in-service corrosion monitoring is not as widespread as the use of polarization resistance. However, this approach can be valuable in some cases, particularly where an alloy could show both active and passive corrosion behavior in a given process stream. For example, stainless steels can provide excellent service, as long as they remain passive. However, if an upset occurs that would introduce either chlorides or reducing agents into the process stream, stainless alloys may become active and may exhibit excessive corrosion rates. Corrosion-potential measurements would indicate the development of active corrosion, and they may be coupled with polarization-resistance measurements as additional confirmation of high corrosion rates (Ref 8). The success of corrosion-potential measurements depends on the long-term stable performance of a standard reference electrode. Such electrodes have been developed for continuous pH monitoring of process streams, and their application for measuring corrosion potentials is straightforward. However, the conditions of temperature, pressure, electrolyte composition, pH, and other possible variables can limit the applications of these electrodes for corrosion-monitoring service.

Reference cited in this section 8. S.W. Dean, Electrochemical Methods for Corrosion Testing in the Process Industries, Electrochemical Corrosion Testing with Special Consideration of Practical Applications, E. Heintz, J.C. Rowlands, and F. Mansfeld, Ed., Proceedings of International Workshop, (Ferrara, Italy), DECHEMA, 1986, p 1–15


Alternating Current Impedance Measurements Alternating current (ac) impedance-polarizing techniques are being increasingly used in electrochemical evaluation. In principle, the same probe elements used for polarization-resistance measurements can be used for ac impedance studies. As shown in Fig. 6, the ac impedance approach permits the separation and independent analysis of the resistive and capacitive elements of the electrochemical corrosion reaction (Ref 26). The method is particularly useful for high- resistivity electrolytes, such as steam condensate. In these cases, polarizationresistance measurements provide readings that are erroneously low, because the bulk resistance of the electrochemical cell adds to the measured resistance. This problem is minimized with the ac approach. ASTM G 96 (Ref 14) provides a very comprehensive discussion of this problem and gives practical guidance on dealing with it.

Fig. 6 Typical plot of ac impedance data in terms of real and imaginary resistance. The numbers on the curve indicate the frequency in hertz. Source: Ref 7 A sophisticated ac frequency generator and analyzer system is necessary to obtain results such as those shown in Fig. 6. Some investigators have suggested using only two frequencies to characterize ac impedance behavior (Ref 27). Although this minimizes the number of measurements, it does not reveal the details of the behavior. It is necessary to be familiar with the approach in order to interpret the results of ac impedance studies. The specimen designs used in commercial polarization-resistance probes have not been optimized for ac analysis. For this reason, a full frequency-response curve is desirable in order to verify that the behavior can be analyzed by conventional techniques, although it may take several hours to obtain a complete frequency-response curve. Thus, more development is necessary before the ac impedance technique achieves widespread acceptance for in-service corrosion monitoring. Another drawback is that the Stearn-Geary constant (Ref 25) or Tafel slopes must be known in order to convert ac impedance data into corrosion-rate information.

References cited in this section 7. S.W. Dean, Overview of Corrosion Monitoring in Modern Industrial Plants, Corrosion Monitoring in Industrial Plants Using Nondestructive Testing and Electrochemical Methods, STP 908, G.C. Moran and P. Labine, Ed., American Society for Testing and Materials, 1986, p 197–220 14. “Standard Guide for On-Line Monitoring of Corrosion in Plant Equipment (Electrical and Electrochemical Methods),” G 96, Annual Book of ASTM Standards, Vol 03.02, ASTM International 25. M. Stern and R.M. Roth, J. Electrochem. Soc., Vol 104, 1958, p 440t 26. S.W. Dean, in Electrochemical Techniques for Corrosion, R. Baboian, Ed., National Association of Corrosion Engineers, 1977, p 52–60 27. S. Haruyama and T. Tsuru, in Electrochemical Corrosion Testing, STP 727, F. Mansfeld and U. Bertocci, Ed., American Society for Testing and Materials, 1981, p 167–186


Electrochemical Noise Electrochemical noise refers to the natural fluctuations of either current or potential of a corroding specimen (Ref 28). If a specimen is electrically isolated and allowed to corrode, there are fluctuations of the corrosion potential as the corrosion proceeds. Likewise, if a corroding specimen is held at constant potential, fluctuations occur in the current required to maintain that potential. In one type of setup, three identical specimens are exposed to the environment. The potential between one of the specimens and the other two is monitored, while the current between the other two is also monitored by a zero- resistance ammeter. These signals are subsequently analyzed to obtain a standard deviation of the current noise and a standard deviation of the potential noise. The ratio of the potential standard deviation to the current standard deviation is believed to be equal to the polarization resistance, and this can then be converted to a corrosion rate through the Stearn-Geary relationship (Ref 25). This approach has appealed to some investigators, because it does not involve any external polarization of the specimens and does not require the use of reference electrodes. Localized corrosion produces unique noise signals, and this has attracted attention because it gives this approach a chance of detecting these forms of corrosion while they are initiating. Pitting, for example, has been characterized by the ratio of the current standard deviation to the average corrosion current. Other analyses of the noise characteristics have been proposed but not universally accepted.

References cited in this section 25. M. Stern and R.M. Roth, J. Electrochem. Soc., Vol 104, 1958, p 440t 28. J.R. Scully, Eletrochemical Tests, Corrosion Tests and Standards, Manual 20, ASTM International, 1995, p 75–90


Hydrogen Probe The concept of the hydrogen probe is based on the fact that one of the cathodic reaction products in nonoxidizing acidic systems is hydrogen. The hydrogen atoms thus generated diffuse through the thickness of the vessel and are liberated at the exterior surface. A hydrogen probe—in one variation, the probe is a piece of palladium foil—is applied to the exterior surface. The probe serves as collector and catalyst for the subsequent oxidation of the hydrogen (Ref 26). The cell is attached to the palladium foil. The current employed in the oxidation reaction is supplied by a stainless steel working electrode (Fig. 7). The electrolyte is 90% sulfuric acid. The current that is required to oxidize the hydrogen is measured; it is directly proportional to the corrosion rate of the interior of the vessel.

Fig. 7 Design of palladium foil hydrogen patch probe. Source: Ref 27 Hydrogen-probe analysis measures the corrosion rate, unlike ultrasonic thickness measurements and other techniques that measure remaining wall thickness. However, hydrogen-probe analysis is limited to systems in which the temperature is close to ambient and the diffusion rate of hydrogen is high. Gas pipeline service is the most common application. In this case, corrosion can occur when hydrogen sulfide, water, and sometimes carbon dioxide (CO2) are present. This approach has another variation that consists of simply attaching a chamber to the exterior of the pipe and monitoring hydrogen liberation through increasing pressure (Ref 2). One advantage of exterior hydrogen monitoring is that it does not require penetration of the pipe wall in order to obtain the corrosion rate. It is assumed in this method that all of the hydrogen liberated in the corrosion reaction diffuses through the steel vessel wall instead of being liberated as hydrogen gas at the surface, and this is true when hydrogen recombination poisons, such as hydrogen sulfide (H 2S), are present. Commercial devices are available for corrosion monitoring by this technique, although it is questionable whether such devices could be positioned and allowed to operate unattended for extended periods of time. In addition, these units have all of the problems associated with any electrochemical measuring device, namely the need for complex electronic equipment and wiring and the need for operators and installers with a sensitivity to the requirements of such equipment. Also, this method is, in practice, limited to steel, which has a high hydrogen diffusivity and low solubility of hydrogen.

Exterior hydrogen monitoring does not supply a quantitative measurement of hydrogen damage. However, it does provide a means of measuring the relative severity of corrosion damage and of evaluating the effects of process changes on the severity of corrosion.

References cited in this section 2. C.P. Dillon, A.S. Krisher, and H. Wissenberg, Plant Corrosion Tests, Handbook on Corrosion Testing and Evaluation, W.H. Ailor, Ed., John Wiley & Sons, 1971, p 599–615 26. S.W. Dean, in Electrochemical Techniques for Corrosion, R. Baboian, Ed., National Association of Corrosion Engineers, 1977, p 52–60 27. S. Haruyama and T. Tsuru, in Electrochemical Corrosion Testing, STP 727, F. Mansfeld and U. Bertocci, Ed., American Society for Testing and Materials, 1981, p 167–186


Analysis of Process Streams Another useful in-service corrosion-monitoring technique is analysis of the process streams for the presence of corrosion products. This straight-forward approach usually does not require the installation of specialized equipment. For example, process streams from the bottom of the distillation column can be routinely sampled, and atomic absorption analysis techniques can be used to determine such heavy metals as iron, nickel, and chromium at very low levels. The concentration of such impurities is then directly proportional to the corrosion rate multiplied by the area of metal corroding, if the only source of metal ions is corrosion, and if the corrosion products are not precipitating. One problem is that the corroding area may not be known with certainty; therefore, the results are relative. However, they do help to determine whether conditions have improved.


Sentry Holes In this monitoring technique, small sentry holes are drilled into the outside of a vessel or pipe at areas that are considered to be particularly susceptible to corrosion. The holes are drilled to the pressure-design thickness. Thus, when corrosion has almost consumed the corrosion allowance, the appearance of a small leak indicates that action must be taken to prevent a major failure. Sentry holes may be threaded or may have nipples attached

to facilitate plugging. Nondestructive testing is frequently performed in the area near the leak to determine the extent of the damage before repair or shutdown. These holes are required in reinforcing pads applied to support nozzles in pressure vessels. In these cases, leakage from the hole signals corrosion or cracking of the nozzle weld and is an important safety feature of the design.


Side-Stream (Bypass) Loops It is sometimes advantageous to operate a side-stream loop to determine the effect of chemical process changes without making these changes in the entire process system. This provides the advantage of an in-service corrosion test with product streams, yet permits evaluation of the effects of additives, inhibitors, and other changes in the environment without affecting the main process stream. In the simplest case, a side stream from a major piece of equipment is run through a small tank or a widened section of the line in order to permit changes in the product stream and insertion of corrosion-monitoring devices. In such a case, the effluent from the side stream is usually discarded. In a more complicated installation, the entire cycle of the operation may be duplicated on a small scale using a portion of the mainstream. Obviously, a more sophisticated side-stream apparatus can approach the complexity and the expense of a pilot plant.


Strategies in Corrosion Monitoring Monitoring Locations. A principal part of any corrosion-monitoring program is deciding where to locate the corrosion-monitoring devices. Because corrosion will probably not occur uniformly throughout the plant, it is desirable in any corrosion-monitoring program to find sites at which the highest corrosion rates will be experienced. The problems involved in developing corrosion-monitoring programs for a plant are illustrated in the example of a distillation column. The most logical points for corrosion monitoring in a distillation column are the feed point, the overhead product receiver, and the reboiler or bottoms product line. These points are the locations at which the highest and lowest temperatures are encountered, as well as the points at which the most and least volatile products are concentrated. However, these points are usually not the locations of the most severe corrosion. The species causing the corrosion often concentrates at an intermediate point in the column because of chemical changes within the column. Therefore, if there is a possibility of concentrating a corrosive species within the distillation column, several monitoring points would be required throughout the column for the corrosion-

monitoring program to be comprehensive (Fig. 8). A monitoring program can be restricted to the most corrosive location within the column, once this area has been identified (Ref 29).

Fig. 8 Distillation column showing preferred locations of monitoring probes or other devices. Source: Ref 7 Another problem with distillation columns is that the liquid on trays tends to be frothy. This creates difficulties for electrochemical methods. One solution to this problem is to install bypass loops that remove liquid from the column, pass it over the corrosion-monitoring probes, and reinject it at a lower point in the column (Ref 2). This practice avoids the problem of foam and froth and provides a more controlled flow rate over the corrosionmonitoring equipment. Use of a bypass loop also allows removal of liquid samples at times of high corrosion rates (Fig. 9).

Fig. 9 Use of a side-stream (bypass) loop to monitor corrosion in a distillation column. Source: Ref 7 The process stream can be sampled in locations other than distillation columns. A sample tap can be helpful in conjunction with polarization-resistance devices. The alarm on polarization-resistance monitoring equipment can be used to signal the need to remove samples. This is particularly desirable in the case of pilot plant operation, in which wide variations in processing conditions are encountered. It is also helpful in plants that produce different products in the same equipment. High-velocity gas streams in pipes may cause problems with conventional monitoring systems. In this case, the presence of an aqueous phase is usually restricted to a thin layer on the surface of the pipe. This condition is known as annular flow. A probe that protrudes into the pipe may miss the liquid layer that is present only close to the pipe wall. A flush-mounted surface probe can be used in such cases. This probe permits the measurement of polarization resistance in order to estimate the corrosion rate of the pipe wall (Ref 30). Probe location is also critical in storage tanks containing nonaqueous liquids. The most corrosive location in these tanks may be at the liquid level, if the liquid in the tank has a density exceeding that of water. In this case, the corrosion- monitoring probe should be mounted on a floating platform. In this way, the probe can detect the presence of a corrosive aqueous phase. However, when the liquid stored in the tank is less dense than the water, the corrosion-monitoring device should probably be positioned at the bottom of the tank. Ultrasonic Thickness Data Analysis. Ultrasonic thickness measurements are often obtained during turnaround periods when the plant is not operating. A well-designed program provides a series of thickness measurements at specified locations throughout the plant. Assuming that the original thickness of a vessel was constant, a diagram similar to that shown in Fig. 10 can be drawn. Figure 10 shows the inverse relationship between the average corrosion rate and the remaining thickness. It is then possible to estimate the remaining life of the vessel from the maximum corrosion rate and the remaining corrosion allowance. In cases in which corrosion rate varies randomly, the highest corrosion rate and the smallest remaining thickness can be estimated by using statistical analysis. This approach provides a rational method of estimating remaining life when faced with a quantity of thickness readings from an ultrasonic thickness survey.

Fig. 10 Typical plot of metal thickness remaining versus calculated corrosion rate obtained from ultrasonic thickness data. Statistical analysis can be used to estimate the remaining corrosion allowance in terms of standard deviation σ. Source: Ref 7 Redundancy is also important in designing corrosion-monitoring programs. The use of at least two different types of corrosion-monitoring devices at any location is often desirable. For example, the use of an electricalresistance probe with a polarization-resistance probe allows measurement of both instantaneous corrosion rates and an average corrosion rate. The data thus obtained can be correlated, and this is very helpful in identifying spurious or inaccurate readings. In another approach, the polarization-resistance probe is weighed before and after the test in order to correlate mass loss with the average corrosion rate that the probe suffered. There is reason to expect that the electrochemical value is in error if the average corrosion rate and the mass loss of the probe do not agree. Also, to obtain time-average corrosion-rate values, independent coupons can be installed together with polarizationresistance probes. A variety of corrosion-monitoring approaches must be used when designing pilot or demonstration plants. Coupon tests can be very helpful in selecting optimal materials for processes based on pilot-plant experience. Polarization-resistance monitoring is very useful for determining whether certain processing conditions cause corrosive situations to develop. Because the corrosion mechanisms are often not well understood, and because the result of erroneous information can be serious overdesign or exposure to unexpected hazards, redundancy in the design of such monitoring systems is important in pilot plants. Other Issues. A primary difficulty with corrosion-monitoring equipment is the need to install wires from the probes to the control room or to the instruments and data-storage systems. Hardwire systems are usually more expensive than the probes and electronic instruments. In addition, wiring systems are often sources of problems due to breakage, moisture entry, and connection difficulties. One option is the use of devices to transform the data to coded citizens' band (CB) radio signals. These small, battery- operated units can be installed at probe locations, and they provide the desired information to the base station on command. The base station would be located in the control room. Acceptance of this approach may increase as such systems become more available and more reliable. The problem of leaks is another important issue in corrosion-monitoring systems. Leaks can cause hazards and can shut down operating plants. The corrosion-monitoring system must be installed such that leaks from the probe can be handled with minimal interruption. A device that penetrates the wall of the vessel may leak; therefore, it is essential to have contingency plans for dealing with leaks before the devices are installed. A related problem concerns packing glands on pressure vessels that have removable devices. The pressure within a system exerts a force on any removable device. Therefore, it is important to prevent the device from being blown out, which might injure personnel attempting to remove the device or would expose such personnel to the fluid within the vessel. For this reason, restraining rods, chains, and flush and drain ports must be used.

References cited in this section 2. C.P. Dillon, A.S. Krisher, and H. Wissenberg, Plant Corrosion Tests, Handbook on Corrosion Testing and Evaluation, W.H. Ailor, Ed., John Wiley & Sons, 1971, p 599–615 7. S.W. Dean, Overview of Corrosion Monitoring in Modern Industrial Plants, Corrosion Monitoring in Industrial Plants Using Nondestructive Testing and Electrochemical Methods, STP 908, G.C. Moran and P. Labine, Ed., American Society for Testing and Materials, 1986, p 197–220 29. R. Martin and E.C. French, “Corrosion Monitoring in Sour Systems Using Electrochemical Hydrogen Potential Probes,” Paper 6657, presented at the Sour Gas Symposium of the Society of Petroleum Engineers (Dallas, TX), American Institute of Mining, Metallurgical and Petroleum Engineers, 1977 30. C.G. Arnold and D.R. Hixson, Mater. Perform., Vol 21 (No. 4), April 1982, p 21–24


Interpretation and Reporting In-service monitoring, more than any other type of corrosion testing, requires the utmost skill in interpretation and reporting. Important economic decisions are often based on the test results. Although there are a number of standards that provide guidelines for certain procedures, there is no standard that is comprehensive. To plan an appropriate test program, the investigator must know or have good advice on both the chemistry and the mechanics of the processes involved; that is, the investigator must understand the entire corrosion system. There must be a strong emphasis on the strategy of the program and a searching analysis of the test results. It is important to prepare detailed records of what was done as part of the experiment, and it is also important to document any unplanned changes that occur in the process stream or the equipment during the investigation. Without a valid interpretation and effective (timely) reporting, the price of the work can be significantly greater than the cost of time and materials. However, the consequences of corrosion failures go beyond additional costs. Also involved are personal safety risks (and liability), hazard potentials, process reliability, and product quality and pollution problems.


Acknowledgments

S.W. Dean reviewed and made substantial revisions to the original article, “In-Service Monitoring,” by S.W. Dean and D.O. Sprowls, Metals Handbook, 9th edition, Vol 13, 1987, p 197. Portions of the original article were adapted with permission from S.W. Dean, Overview of Corrosion Monitoring in Modern Industrial Plants, Corrosion Monitoring in Industrial Plants Using Non-Destructive Testing and Electrochemical Methods, STP 908, G.C. Moran and P. Labine, Ed., American Society for Testing and Materials, 1986, p 197–220.


References 1. G.H. Koch, M.P.H. Brongers, N.G. Thompson, Y.P. Virmani, and J.H. Payer, “Corrosion Cost and Preventive Strategies in the United States,” FHWA-RD-01-156, Federal Highway Administration, U.S. Department of Transportation, Washington D.C., March 2002 2. C.P. Dillon, A.S. Krisher, and H. Wissenberg, Plant Corrosion Tests, Handbook on Corrosion Testing and Evaluation, W.H. Ailor, Ed., John Wiley & Sons, 1971, p 599–615 3. A.S. Couper, Methods in Oil-Refining Industries, Handbook on Corrosion Testing and Evaluation, W.H. Ailor, Ed., John Wiley & Sons, 1971, p 617–624 4. “Laboratory Corrosion Testing of Metals for the Process Industries,” NACE TM-01-69, National Association of Corrosion Engineers, 1976 5. J.N. Wanklyn, Corrosion Monitoring in the Oil and Petrochemical and Process Industries, J. Wanklyn, Ed., Oyez Scientific and Technical Service Ltd., 1982 6. Corrosion Monitoring in Industrial Plants Using Nondestructive Testing and Electrochemical Methods, STP 908, G.C. Moran and P. Labine, Ed., American Society for Testing and Materials, 1986 7. S.W. Dean, Overview of Corrosion Monitoring in Modern Industrial Plants, Corrosion Monitoring in Industrial Plants Using Nondestructive Testing and Electrochemical Methods, STP 908, G.C. Moran and P. Labine, Ed., American Society for Testing and Materials, 1986, p 197–220 8. S.W. Dean, Electrochemical Methods for Corrosion Testing in the Process Industries, Electrochemical Corrosion Testing with Special Consideration of Practical Applications, E. Heintz, J.C. Rowlands, and F. Mansfeld, Ed., Proceedings of International Workshop, (Ferrara, Italy), DECHEMA, 1986, p 1–15 9. Process Industries Corrosion—Theory and Practice, National Association of Corrosion Engineers, 1986 10. B.J. Moniz, Field Coupon Corrosion Testing, Process Industries Corrosion—Theory and Practice, National Association of Corrosion Engineers, 1986 11. S.J. Pikul, T.S. Lee, and D.B. Anderson, “Field Corrosion Tests for Materials Evaluation in Chemical Process Industries,” Paper 119, presented at Corrosion/86, National Association of Corrosion Engineers, 1986

12. A. Perkins, Industrial Applications, Corrosion Tests and Standards, Manual 20, ASTM International, 1995, p 143–148 13. “Standard Method for Conducting Corrosion Coupon Tests in Plant Equipment,” G 4, Annual Book of ASTM Standards, Vol 03.02, ASTM International 14. “Standard Guide for On-Line Monitoring of Corrosion in Plant Equipment (Electrical and Electrochemical Methods),” G 96, Annual Book of ASTM Standards, Vol 03.02, ASTM International 15. “Standard Guide for Conducting and Evaluating Galvanic Corrosion Tests in Electrolytes,” G 71, Annual Book of ASTM Standards, Vol 03.02, ASTM International 16. W.D. France, Jr., Crevice Corrosion of Metals, Localized Corrosion—Cause of Metal Failure, STP 519, M. Henthorne, Ed., American Society for Testing and Materials, 1972, p 164–200 17. “Standard Guide for Crevice Corrosion Testing of Iron-Base and Nickel-Base Stainless Alloys in Seawater and Other Chloride- Containing Aqueous Environments,” G 78, Annual Book of ASTM Standards, Vol 03.02, ASTM International 18. “Standard Test Methods for Pitting and Crevice Corrosion Resistance of Stainless Steels and Related Alloys by Use of Ferric Chloride Solution,” G 48, Annual Book of ASTM Standards, Vol 03.02, ASTM International 19. “Standard Practice for Making and Using U- Bend Stress-Corrosion Test Specimens,” G 30, Annual Book of ASTM Standards, Vol 03.02, ASTM International 20. “Standard Practice for Making and Using C- Ring Stress-Corrosion Test Specimens,” G 38, Annual Book of ASTM Standards, Vol 0302, ASTM International 21. “Standard Practice for Preparation of Stress- Corrosion Test Specimens for Weldments,” G 58, Annual Book of ASTM Standards, Vol 03.02, ASTM International 22. “Standard Practice for Determining the Susceptibility to Intergranular Corrosion of 5xxx Series Aluminum Alloys by Weight Loss After Exposure to Nitric Acid (NAWLT Test),” G 67, Annual Book of ASTM Standards, American Society for Testing and Materials 23. E.H. Dix, Jr., W.A. Anderson, and M.B. Shumaker, Influence of Service Temperature on the Resistance of Wrought Aluminum-Magnesium Alloys to Corrosion, Corrosion, Vol 15 (No. 2), 1959, p 55t–62t 24. “Standard Practice for Preparing, Cleaning, and Evaluating Corrosion Test Specimens,” G 1, Annual Book of ASTM Standards, Vol 03.02, ASTM International 25. M. Stern and R.M. Roth, J. Electrochem. Soc., Vol 104, 1958, p 440t 26. S.W. Dean, in Electrochemical Techniques for Corrosion, R. Baboian, Ed., National Association of Corrosion Engineers, 1977, p 52–60 27. S. Haruyama and T. Tsuru, in Electrochemical Corrosion Testing, STP 727, F. Mansfeld and U. Bertocci, Ed., American Society for Testing and Materials, 1981, p 167–186 28. J.R. Scully, Eletrochemical Tests, Corrosion Tests and Standards, Manual 20, ASTM International, 1995, p 75–90

29. R. Martin and E.C. French, “Corrosion Monitoring in Sour Systems Using Electrochemical Hydrogen Potential Probes,” Paper 6657, presented at the Sour Gas Symposium of the Society of Petroleum Engineers (Dallas, TX), American Institute of Mining, Metallurgical and Petroleum Engineers, 1977 30. C.G. Arnold and D.R. Hixson, Mater. Perform., Vol 21 (No. 4), April 1982, p 21–24 31. E.C. French, Oil Gas J., Vol 17, 1975, p 82–88

B. Phull, Evaluating Uniform Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 542–544

Evaluating Uniform Corrosion Revised by Bopinder Phull, Consultant

Introduction UNIFORM CORROSION refers to attack on an exposed metal surface that results in homogeneous thickness loss. In other words, the metal loss is spread evenly over the surface rather than being localized as, for example, pitting or stress- corrosion cracking. Uniform corrosion is also known as general corrosion. Practical examples of uniform or general corrosion include dissolution of iron or zinc in dilute sulfuric acid, aluminum in sodium hydroxide solution, bare or galvanized steel in atmospheric environments, and so on. In terms of electrochemistry, the metal undergoing uniform corrosion can be regarded as a single electrode, which supports anodic and cathodic reactions concurrently. Furthermore, the anodic and cathodic sites can be considered as local with respect to each other, approximately equal in area, and fluctuating randomly in a stochastic manner over the metal surface. Corrosion occurs by local cell action, that is, metal dissolution of the anodic sites due to oxidation (electron loss), driven by charge-transfer reduction reaction(s) at the adjacent cathodic sites (electron consumption). Uniform corrosion can be mitigated by materials selection, coatings, inhibitors, cathodic protection, and in specific cases anodic protection. Another practical means of addressing uniform corrosion in many practical applications is to build in a corrosion allowance at the design stage. The most appropriate corrosion control method will depend on specific metal-environment considerations, application criticality, and economics. Testing for uniform corrosion and its evaluation are relatively simple. Appropriate test specimens (also often referred to as corrosion coupons) are exposed to the environment of interest, and the metal loss is determined after a prescribed period by, for example, mass loss or measurement of thickness change. Convenience, practicality, safety considerations, and economics will typically dictate whether testing is performed in the laboratory and/or in actual service conditions. Laboratory testing usually consists of small specimens immersed in relatively small volumes of simulated or actual service environments. The advantage of this type of testing is that the tests can be controlled closely; thus, unintentional disturbances that could occur during in-plant tests can be avoided. These tests are usually designed to simulate expected service conditions. It should be recognized that the objective of testing is to determine the type and extent of corrosion incurred. There is no advance assurance that the test specimens will undergo uniform corrosion. Service tests (field, in-plant, or on-line tests) involve placing test samples in actual service conditions for evaluation. The test specimens need to be mounted securely and are usually electrically isolated from each other and plant equipment to preclude galvanic interactions— unless the objective of the test is to investigate such effects. The advantage of service testing is that the test specimens are exposed to the actual service environment. However, in-plant testing is also fraught with limitations. For instance, it may be difficult to subject the test specimens to the same flow, pressure, or heat-transfer conditions as the actual plant equipment, such as in piping, pumps, valves, and condensers. Although techniques are available commercially for on-line

insertion and retrieval of test specimens, that is, without plant shutdown, they are not used routinely in all industries because of cost.

B. Phull, Evaluating Uniform Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 542–544 Evaluating Uniform Corrosion Revised by Bopinder Phull, Consultant

Mass-Loss Tests For engineering applications, uniform corrosion rates are reported as loss of metal thickness per unit time. Corrosion rates are often calculated from mass-loss data. Mass loss is the difference between the original mass of the specimen and the mass after a prescribed or known exposure period. Before determining the mass loss after exposure, it is important to clean the specimens properly to remove corrosion product adhering to the surface. After cleaning, it is essential to examine the test specimens visually. Calculation of corrosion rates from mass-loss data will yield meaningful results for use in design only if the specimens have suffered uniform attack. Conversely, computation of corrosion rates from mass loss for specimens that have predominately incurred localized corrosion will give erroneous information to the designer or end-user. It may be possible to measure the thickness changes in the test specimens or equipment directly, for example, by using calipers, micrometers, or ultrasonic gages. Ultrasonic thickness measurements are particularly suited to scheduled inspection of actual plant equipment such as piping, vessels, tanks, pump casings, and valve bodies. Uniform corrosion rates are usually expressed as thickness loss per unit time, for example, micrometers/year (μm/y), millimeters per year (mm/y), and mils per year (mpy; where 1 mil = 25.4 μm). Specimen Preparation. The first step in testing is the selection and preparation of suitable test specimens. Clear and complete documentation of the test material is very important. This should include the chemical composition as well as the metallurgical condition of the specimens. The specimens are usually saw-cut from larger stock and the surfaces and edges ground and deburred. It is important to minimize introduction of stresses or metallurgical transformations (e.g., due to overheating) during specimen preparation. Laser cutting, while rapid and convenient, may introduce microstructural changes that make the heat-affected areas, for example, specimen edges or mounting holes, more or less prone to corrosion, depending on the material and test environment. Each sample should be clearly identified before testing. This is commonly done with identification codes that are stenciled (stamped or engraved) on the specimens. Stamped codes introduce stresses where cracking can occur in susceptible materials. Similarly, codes made by laser stenciling or electrical discharge machining (EDM) may increase corrosion propensity of associated areas. It is important to prepare a map sheet so that test specimens can still be identified even if the identification codes are obliterated by corrosion during the test. The size and shape of the test specimens are often specified in such a manner as to make them easy to handle and allow for ease of surface preparation, cleaning, and weighing. Specimens 1.6 to 6.4 mm ( to in.) thick and a few inches square are not uncommon. While it may be desirable to protect specimen edges for some materials, this is not always practical in corrosion tests. Before exposure to the corrosive medium, the test specimen surface needs to be prepared and cleaned. The surface can be ground, grit or glass-bead blasted, or pickled. Ideally, the surface finish of the test specimen should be the same as that of the metal or alloy in service. Unfortunately, this is not always possible because of the variations in surface condition produced during fabrication. However, surface preparation of the test specimens ensures a standard surface condition during testing. Usually, sufficient material is removed to reach the bulk composition/ structure of the metal or alloy; that is, surface- depleted or -enriched layers nonrepresentative of bulk material are removed, unless influence of these variables is part of the investigation. After surface preparation, the specimens are degreased with a solvent such as acetone, dried, weighed, and either immediately exposed to the test medium or stored in a desiccator until test commencement. For most laboratory tests, specimens should ideally be weighed to a

precision of ±0.1 mg. Ideally, specimens with a large surface-area-to-volume ratio are preferred for accurate mass-loss determination, especially when the expected corrosion rate is low. Practical guidelines for preparing, cleaning, and evaluating corrosion test specimens are provided in ASTM G 1 (Ref 1). Test Methods. The method and apparatus used to expose the test specimens to the test environment are very important and should meet the following guidelines: • • • • •

The corrosive medium should have easy access to the test specimens. The specimens should be well supported. The specimens should be electrically insulated from all other metals in the system. The specimens should be completely immersed, unless other effects are being studied. The specimens should be readily accessible.

Practical guidelines for laboratory immersion corrosion testing of metals are provided in ASTM G 31 (Ref 2) and NACE International TM0169-2000 (Ref 3). It is important to plan the experimental technique so that the test specimens are removed and the mass loss determined after appropriate exposure duration. The corrosion rate may initially increase or decrease and eventually become constant with time. Exposure of many replicate specimens is usually necessary with the mass- loss test method to determine corrosion rate variations as a function of time. The following very rough rule of thumb suggested for test duration (in hours) is cited in Refs 2 and 3: (Eq 1) In addition to the sequential removal of specimens with time, a planned interval testing schedule can be used to obtain more information on the initial corrosion rates and on accumulated effects (see ASTM G 31, Ref 2). Planned interval testing involves exposing identical specimens to the test solution at various intervals during the test. By evaluating the results, initial corrosion rates and changes in solution corrosivity can be determined. Planned interval testing is an excellent example of good experimental planning. References 2 and 3 also include recommendations for minimum solution volume to use in the test per unit surface area of exposed specimen (e.g., 20 mL/cm2, or 125 mL/in2). This is to avoid depletion of corrodents and hence significant changes in solution corrosivity during the test. Test Variables. Important variables that must be addressed during experimental planning include gases, temperature, solution volume and flow, and exposure time. The presence of dissolved gases, for example, O 2, CO2, H2S, SO2, and so on, in the test solution can have a dramatic effect on the rate of corrosion. Dissolved gases are usually supplied by bubbling them through the solution. Running the test under pressure or partial vacuum will alter the dissolved gas concentration. If degassing is required, purified nitrogen or argon can be bubbled through the solution to displace dissolved gases. The influence of temperature can be complicated in that it affects gas solubility, pH, corrosion product formation, and other parameters. Therefore, the temperature of the test should be monitored and controlled as necessary during the corrosion test. The flow velocity of the solution also plays an important role in determining the corrosion rates. Corrosion rates may increase or decrease as the fluid velocity near the sample is increased. In any case, flow characteristics that represent the condition in service should be established during testing. This is often not a simple task, because the hydrodynamic conditions in service and their effect on the corrosion rate can be complex. Rotating disks and cylinders and jet-impingement apparati have been used to study corrosion rates under controlled hydrodynamic conditions (Ref 4, 5, 6, 7). The type of exposure is also important in test design; total, partial, or intermittent exposure can change the mode and rate of corrosion. For immersion testing, all test specimens should be immersed at approximately the same depth in the test environment. Partially immersed samples can be used to study waterline corrosion. Samples can also be tested in the vapor phase. Posttest Sample Cleaning, Data Acquisition, and Reporting. After the test specimens are removed from the exposure environment, they are cleaned to remove any corrosion products. The corrosion products must be removed with minimal additional corrosion of the test specimens. Sometimes, the corrosion products are not removed, and the mass gain is used to calculate corrosion rate. In principle, this approach has some appeal, because the specimen can be returned to the test after weighing (i.e., without removing corrosion products).

However, it assumes that all of the corrosion products formed from the inception of the test remain on the surface and that their stoichiometry is well known. Mass-gain testing has been used in high- temperature oxidation, gaseous, and atmospheric environments. However, the more common method is to remove the corrosion products by a combination of careful chemical and mechanical cleaning before determining mass loss. Detailed procedures and solutions for cleaning specimens after exposure are given in ASTM G 1 (Ref 1), including how to make corrections for mass loss incurred during cleaning. It is essential to examine the test specimens visually to verify or characterize the type of corrosion attack suffered. This can generally be done with the unaided eye and does not usually require the aid of a microscope. If corrosion is uniform, the corrosion rate can be calculated from the mass-loss data as follows: Corrosion rate = (K × W)/(A × t × ρ)

(Eq 2)

where K is a constant, W is mass loss, A is exposed specimen area, t is exposure time, and ρ is material density. Table 1 shows the various corrosion rate units and corresponding values of the constant K in Eq 2 when W is in grams, A is in cm2, t is in hours, and ρ is in g/cm3 (Ref 1, 2). ASTM G 1 also contains material density data for common metals and alloys. It must be reiterated that corrosion rates that reflect material-thinning rates should not be calculated from mass-loss data when corrosion attack is primarily localized, because they will lead to erroneous interpretation and conclusions. For example, calculation of corrosion rate from mass-loss data alone may indicate a negligible uniform corrosion rate for a given material, whereas, in reality, the subject test specimen may have been completely perforated by a few small-diameter pits. In other words, the low uniform corrosion rate would provide a false sense of confidence in the corrosion performance that would be contradicted by unexpected failure in service (e.g., heat-exchanger tubing). In the case of uniform and nonuniform corrosion, units such as g/m2 · h or mdd (see Table 1) can be used for making relative comparisons of corrosion performance, for example, evaluation of solution corrosivity or determining efficacy of corrosion mitigation methods. However, the data should not be converted to thinning rates (e.g., mpy or mm/y) unless the corrosion is confirmed to be uniform. Table 1 Values for constant K in the corrosion rate equation (Eq 2) Corrosion rate units desired Constant K in corrosion rate equation 3.45 × 106 Mils per year (mpy) 3.45 × 103 Inches per year (ipy) 8.76 × 107 Micrometers per year (μm/y) 8.76 × 104 Millimeters per year (mm/y) 2 1 × 104 × ρ Grams per square meter per hour (g/m · h) Milligrams per square decimeter per day (mdd) 2.40 × 106 × ρ ρ is material density in g/cm3. Source: Ref 1 ASTM G 4 (Ref 8) provides guidelines on conducting corrosion coupon tests in field applications, particularly the use of test racks in plant applications.

References cited in this section 1. “Standard Practice for Preparing, Cleaning, and Evaluating Corrosion Test Specimens,” G 1, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Standards and Testing, 2002 2. “Standard Practice for Laboratory Immersion Corrosion Testing of Metals,” G 31, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Standards and Testing, 2002 3. “Standard Test Method: Laboratory Corrosion Testing of Metals,” TM0169-2000, NACE International, 2000 4. J.S. Newman, Electrochemical Systems, Prentice-Hall, 1973

5. D.C. Silverman, Rotating Cylinder Electrode for Velocity Sensitivity Testing, Corrosion, Vol 40, 1984, p 220 6. A.J. Bard and L.R. Faulkner, in Electrochemical Methods: Fundamentals and Applications, J. Wiley & Sons, 1980, p 283–304 7. K.D. Efird, E.J. Wright, J.A. Boros, and T.G. Hailey, Correlation of Steel Corrosion in Pipe Flow with Jet Impingement and Rotating Cylinder Tests, Corrosion, Vol 49, 1993, p 992 8. “Standard Guide for Conducting Coupon Tests in Field Applications,” G 4, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Standards and Testing, 2002


Other Test and Evaluation Methods If visual appearance indicates uniform corrosion on the test specimens, then periodic analysis for dissolved or insoluble corrosion products may be used to calculate corrosion rate (Ref 9). This is generally most meaningful for closed systems; that is, the test solution is not refreshed. This approach assumes that all of the corroded metal enters the test solution does not remain on the surfaces of the test specimens as corrosion product. For an alloy, it is assumed that its constituent elements enter the solution stoichiometrically. For material-environment combinations where hydrogen is evolved as a by-product, the hydrogen can be collected during the test, using a gas burette. Again, if corrosion is uniform, it may be possible to calculate the corrosion rate from the volume of hydrogen evolved, provided the correlation between metal loss and amount of gas liberated is known (Ref 9, 10). If corrosion attack is uniform, useful estimates of corrosion rates can be obtained from electrochemical techniques such as linear polarization resistance (LPR). This technique is described in ASTM G 59 (Ref 11). Basically, a small voltage is applied between two small, identical electrodes made from the material of interest. The electrodes are usually incorporated in a probe, which is immersed in the environment. The current associated with the small voltage perturbation is measured. Polarization resistance is the ratio of applied voltage to the resultant current density and is inversely related to the corrosion rate via a Stern and Geary constant. This technique requires immersion of the probe in an aqueous (preferably conductive) medium. The LPR technique lends itself readily to monitoring uniform corrosion in laboratory tests or in-plant applications (Ref 12). Instrumentation ranges from portable, single-channel, hand-held units to multichannel systems installed in control rooms. Another technique for monitoring uniform corrosion is the electrical resistance (ER) method. The change in the electrical resistance of a sensing element (e.g., wire, strip, or foil) made from the material of interest and exposed to the environment is monitored with a Kelvin- Wheatstone bridge as a function of time. The sensing element is usually incorporated into a suitable probe. The sensing element thins due to corrosion, with a corresponding increase in its electrical resistance. The corrosion rate is directly related to the slope of the curve from plotting resistance change versus time. A major attribute of the ER technique is that, unlike the LPR technique, it does not require an aqueous or conductive medium. Monitoring equipment for laboratory and online corrosion tests using the electrical resistance technique is available commercially (Ref 12).

References cited in this section 9. G. Wranglén, An Introduction to Corrosion and Protection of Metals, Chapman & Hall, 1985, p 229 10. R.L. Martin and E.C. French, Corrosion Monitoring in Sour Systems Using Electrochemical Hydrogen Patch Probes, J. Pet. Technol., Nov 1978, p 1566–1570 11. “Standard Test Method for Conducting Potentiodynamic Polarization Resistance Measurements,” G 59, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Standards and Testing, 2002 12. “Standard Guide for On-Line Monitoring of Corrosion in Plant Equipment (Electrical and Electrochemical Methods),” G 96, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Standards and Testing, 2002


Acknowledgment This article has been adapted in part from C.A. Natalie, Evaluation of Uniform Corrosion, Corrosion, Vol 13, ASM Handbook (formerly Metals Handbook, 9th ed.), ASM International, 1987, p 229–230.


References 1. “Standard Practice for Preparing, Cleaning, and Evaluating Corrosion Test Specimens,” G 1, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Standards and Testing, 2002 2. “Standard Practice for Laboratory Immersion Corrosion Testing of Metals,” G 31, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Standards and Testing, 2002 3. “Standard Test Method: Laboratory Corrosion Testing of Metals,” TM0169-2000, NACE International, 2000

4. J.S. Newman, Electrochemical Systems, Prentice-Hall, 1973 5. D.C. Silverman, Rotating Cylinder Electrode for Velocity Sensitivity Testing, Corrosion, Vol 40, 1984, p 220 6. A.J. Bard and L.R. Faulkner, in Electrochemical Methods: Fundamentals and Applications, J. Wiley & Sons, 1980, p 283–304 7. K.D. Efird, E.J. Wright, J.A. Boros, and T.G. Hailey, Correlation of Steel Corrosion in Pipe Flow with Jet Impingement and Rotating Cylinder Tests, Corrosion, Vol 49, 1993, p 992 8. “Standard Guide for Conducting Coupon Tests in Field Applications,” G 4, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Standards and Testing, 2002 9. G. Wranglén, An Introduction to Corrosion and Protection of Metals, Chapman & Hall, 1985, p 229 10. R.L. Martin and E.C. French, Corrosion Monitoring in Sour Systems Using Electrochemical Hydrogen Patch Probes, J. Pet. Technol., Nov 1978, p 1566–1570 11. “Standard Test Method for Conducting Potentiodynamic Polarization Resistance Measurements,” G 59, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Standards and Testing, 2002 12. “Standard Guide for On-Line Monitoring of Corrosion in Plant Equipment (Electrical and Electrochemical Methods),” G 96, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Standards and Testing, 2002



A.H. Ailor, Ed., Handbook of Corrosion Testing and Evaluation, J. Wiley & Sons, Inc., 1971 R. Baboian, Ed., Manual 20 Corrosion Tests and Standards: Application and Interpretation, ASTM, West Conshohocken, PA, 1995 M. Fontana, Corrosion Engineering, 3rd ed., McGraw-Hill, 1986 P.A. Schweitzer, Corrosion and Corrosion Protection Handbook, Marcel Dekker, 1983 R. Winston Revie, Ed., Uhligh's Corrosion Handbook, 2nd ed., J. Wiley & Sons, Inc., 2000

B. Phull, Evaluating Pitting Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 545–548

Evaluating Pitting Corrosion Revised by Bopinder Phull, Consultant

Introduction PITTING is a form of localized corrosion that is often a concern in applications involving passivating metals and alloys in aggressive environments. Pitting can also occur in nonpassivating alloys with protective coatings or in certain heterogeneous corrosive media. It is a very damaging form of corrosion that is not readily evaluated by the methods used for characterizing metal loss caused by uniform corrosion. It should be recognized that accelerated laboratory test methods allow ranking of relative susceptibility or resistance to pitting corrosion but typically do not provide reliable predictions of time to failure in actual service environments.

B. Phull, Evaluating Pitting Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 545–548 Evaluating Pitting Corrosion Revised by Bopinder Phull, Consultant

Test Methods ASTM G 48 gives procedures for determining the pitting (and crevice) corrosion resistance of stainless steels and related alloys when exposed to an oxidizing chloride environment, namely, 6% ferric chloride (FeCl3) at 22 ± 2 or 50 ± 2 °C (70 ± 3.5 or 120 ± 3.5 °F) (Ref 1). Method A is a 72 h total-immersion test of small coupons that is designed to determine the relative pitting resistance of stainless steels and nickel-base chromium-bearing alloys. Method B is a crevice test under the same exposure conditions, and it can be used to determine both the pitting and crevice corrosion resistance of these alloys. These tests can be used for determining the effects of alloying additions, heat treatments, and surface finishes on pitting and crevice corrosion resistance. Methods A and B are designed to cause breakdown of type 304 stainless steel at room temperature. Methods C and D are used to determine the critical temperature to cause initiation of pitting and crevice corrosion, respectively, in 6% FeCl3 + 1% HCl solution. In each of these methods, the use of FeCl3 solution is justified on the basis that a similar electrolyte chemistry develops within the pits in susceptible ferrous alloys exposed to chloridecontaining service environments. As a result, these tests can give misleading results for alloys intended for chloride-free service environments. ASTM F 746 describes a procedure to determine the resistance to pitting or crevice corrosion of passive metals and alloys from which surgical implants will be produced (Ref 2). This is an electrochemical potentiostatic screening test that is used to rank surgical implant alloys in order of their resistance to localized corrosion in a 0.9% NaCl solution at 37 ± 1 °C (98.6 ± 1.8 °F) (average human body temperature). Using a potentiostat, an anodic potential is applied to a small test specimen in an effort to initiate localized corrosion; the current is then monitored as a function of time at a constant potential. With this method, alloys are ranked in terms of the critical potential for pitting; the more electropositive (more noble) this critical potential is, the more resistant the alloy is to passive film breakdown and to localized corrosion. The method was intentionally designed to cause breakdown of at least one alloy (type 316L stainless steel) that is currently considered acceptable for

surgical implant use. Those alloys that suffer pitting or crevice corrosion during more severe portions of the test do not necessarily suffer localized corrosion when placed within the human body as a surgical implant. ASTM G 61 is an electrochemical test method for conducting cyclic potentiodynamic polarization measurements to determine susceptibility to localized corrosion (Ref 3) of iron-, nickel-, and cobalt-base alloys in chloride environments. A small test specimen is polarized using a potentiostat, a counter electrode, and a reference electrode. Commencing at the free (or open circuit) corrosion potential, a polarization curve is generated by scanning the potential of the test specimen in the anodic direction. (The potential versus log current density plot is usually recorded automatically.) The scan is reversed at some preset criterion (potential or current density). Breakdown potential for pitting, repassivation (protection) potential, and hysteresis behavior (difference between forward and reverse anodic scans) data are acquired from the polarization curve and used to compare the pitting (or crevice) corrosion resistance of test materials. Several different methods for determining the protection potential have been compared (Ref 4). The general interpretation of the results is that pits initiate and grow at potentials more noble than the breakdown (pitting) potential of the passive film. In service environments, this can happen when the corrosion potential shifts sufficiently in the electropositive direction, for example, due to the influence of oxidizing species such as O2, Fe3+, OCl-, and so on. At potentials more active or electronegative than the repassivation (protection) potential, pit initiation and growth are stifled. It should be noted that breakdown potential, repassivation potential, and hysteresis behavior are not material properties; they are empirical values, which are influenced by experimental technique and procedure. Nevertheless, they are useful in ranking alloy performance when nominally the same test procedure is followed. Criticisms of such tests for predicting longterm corrosion behavior are discussed in Ref 5, 6, 7. ASTM G 100 gives a procedure for conducting cyclic galvanostaircase polarization (GSCP) to determine relative susceptibility to pitting (and crevice) corrosion of alloys (Ref 8). In this method, the current is increased and decreased and the potential is monitored. For some materials, specifically aluminum alloy 3003-H14 (UNS A93003), this method gives a more reproducible protection potential value compared to potentiodynamic methods such as ASTM G 61. Although polarization test techniques such as ASTM G 61 or G 100 are used primarily in the laboratory, they can also be used successfully in the field. However, the risk of ground (earth) loops, which can affect measurement capabilities and hence results, is greater for in-plant (field) applications because of the possibility of the large electrical and magnetic fields often associated with industrial equipment. A common ground-loop problem can be circumvented by using a potentiostat that has a floating working (test) electrode terminal instead of a grounded one. Because of the large applied potentials used in these tests, the test specimens are usually irreversibly affected, at least in electrochemical terms. Thus, they are not considered as on-line techniques for routine monitoring of localized corrosion. There are on-line techniques (Ref 9), such as electrochemical linear polarization resistance (LPR) and electrical resistance (ER), but they are more suited to monitoring uniform rather than localized corrosion. Electrochemical noise (EN) is a relatively new technique that is finding increasing use alongside the more traditional electrochemical methods, both in research and the field. Electrochemical potential noise (EPN) involves monitoring small shifts (typically μV to mV range) in the corrosion potential of a test electrode with respect to a suitable reference electrode. Initiation of pits is usually characterized by distinct shifts in potential; repassivation is indicated by potential recovery. Quantitative information on corrosion rate is given by electrochemical current noise (ECN) measurements in which the micro-galvanic current between two nominally identical test specimens made from the material of interest is monitored. References 10, 11, 12 provide additional information on EN techniques.

References cited in this section 1. “Standard Test Methods for Pitting and Crevice Corrosion Resistance of Stainless Steels and Related Alloys by the Use of Ferric Chloride Solution,” G 48, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Testing and Materials, 2002 2. “Standard Test Method for Pitting or Crevice Corrosion of Metallic Surgical Implant Materials,” F 746, Medical Devices, Emergency Medical Services, Vol 13.01, Annual Book of ASTM Standards, American Society for Testing and Materials, 2002

3. “Standard Practice for Conducting Cyclic Potentiodynamic Polarization Measurements for Localized Corrosion,” G 61, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Testing and Materials, 2002 4. M. Hubbell, C. Price, and R. Heidersbach, Crevice and Pitting Corrosion Tests for Stainless Steels: A Comparison of Short- Term Tests With Longer Exposures, Laboratory Corrosion Tests and Standards, STP 866, G.S. Haynes and R. Baboian, Ed., American Society for Testing and Materials, 1985, p 324– 336 5. B.E. Wilde, Critical Appraisal of Some Popular Laboratory Tests for Predicting the Localized Corrosion Resistance of Stainless Alloys in Sea Water, Corrosion, Vol 28 (No. 8), Aug 1972, p 283 6. F.L. LaQue and H.H Uhlig, An Essay on Pitting, Crevice Corrosion and Related Potentials, Mater. Perform., Vol 22 (No. 8), Aug 1983, p 34 7. D. C. Silverman, Practical Corrosion Prediction Using Electrochemical Techniques, Uhlig's Corrosion Handbook, R.W. Revie, Ed., John Wiley & Sons, 2000, p 1179–1225 8. “Standard Test Method for Conducting Cyclic Galvanostaircase Polarization,” G 100, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Testing and Materials, 2002 9. “Standard Guide for On-Line Monitoring of Corrosion in Plant Equipment (Electrical and Electrochemical Methods),” G 96, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Testing and Materials, 2002 10. D.A. Eden, Electrochemical Noise, Uhlig's Corrosion Handbook, R.W. Revie, Ed., John Wiley & Sons, 2000, p 1127–1238 11. J.R. Kearns and J.R. Scully, Ed., Electrochemical Noise Measurement for Corrosion Applications, STP 1277, American Society for Testing and Materials, 1996 12. R.G. Kelly, in Corrosion Tests and Standards: Application and Interpretation, R. Baboian, Ed., American Society for Testing and Materials, 1995, p 172


Examination of Pits ASTM G 46 provides guidance for the identification and examination of pits and evaluating the extent of pitting corrosion (Ref 13). It is important to be able to determine the extent of pitting, either in service applications or in laboratory tests for ranking alloy performance. The following is a summary of the procedures described in detail in Ref 13. Additional guidelines can be found in Ref 14.

Identification and Examination of Pits

Visual examination of the corroded surface can be performed with the unaided eye or a lowpower microscope. The corroded surface is usually photographed, and the size, shape, and density of the pits determined (Fig. 1, 2).

Fig. 1 Variations in the cross-sectional shape of pits. (a) Narrow and deep. (b) Elliptical. (c) Wide and shallow. (d) Subsurface. (e) Undercutting. (f) Shapes determined by microstructural orientation. Source: Ref 13

Fig. 2 Standard rating chart for pits. Source: Ref 13

Metallographic examination can be used to determine whether there is a correlation between pits and microstructure and whether the cavities are true pits or the result of another mechanism, such as intergranular corrosion or dealloying. Nondestructive Inspection. Reference 13 outlines procedures for the nondestructive evaluation of pitted specimens. These include radiographic, electromagnetic, ultrasonic, and dye- penetrant inspection. These methods can be used to locate pits and to provide some information on their size, but they generally cannot detect small pits or differentiate between pits and other types of surface defects.

Extent of Pitting Mass loss is generally not a good indication of the extent of pitting unless uniform corrosion is slight and pitting is fairly severe. If there is significant uniform corrosion, the contribution of pitting to total mass loss is small. Mass loss should not be ignored in every case, however. For example, measurement of mass loss, along with visual comparison of pitted surfaces, may be sufficient to rank the relative resistance of alloys in laboratory tests. Pit depth measurement is generally a better indicator of the extent of localized attack than mass loss. Pit depth measurement can be accomplished by several methods, including metallographic examination, machining, use of a micrometer or depth gage, and microscopic measurements. In the microscopic method, a metallurgical microscope is focused on the lip of the pit and then on the bottom of the pit. The difference between the initial and final readings on the fine-focusing knob or stage vernier is the pit depth.

References cited in this section 13. “Standard Recommended Practice for Examination and Evaluation of Pitting Corrosion,” G 46, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Testing and Materials, 2002 14. F.A. Champion, Corrosion Testing Procedures, 2nd ed., John Wiley & Sons, 1965, p 197


Evaluation of Pitting Pitting can be described in several ways. Reference 13 includes procedures for the use of standard charts, metal penetration, statistical analysis, and loss in mechanical properties to quantify the severity of pitting damage. More than one of these methods can be used. In fact, it is often found that no single method is sufficient by itself. Standard charts such as shown in Fig. 2 can be used to rate pits in terms of density, size, and depth. Columns A and B in Fig. 2 are used to rate the density (that is, the number of pits per unit area on the specimen surface) and the average size of the pits, respectively. Column C rates the average depth of attack. An example of a rating using Fig. 2 might be A-3, B-2, C-3; this rating indicates a density of 5 × 104 pits/m2, an average pit size of 2.0 mm2, and an average pit depth of 1.6 mm. Such charts facilitate communication among those who are familiar with the standard ratings and offer a simple means of storing data for comparison with other test results. However, it can be tedious and time-consuming to

measure all of the pits, and the time spent doing so is usually not justified, because maximum values (especially pit depths) are often more significant than average values. Metal Penetration. In this method, the deepest pits are measured and metal penetration is expressed in terms of the maximum pit depth, an average of, for example, the ten deepest pits, or both. Metal penetration is especially significant when the metal is intended for service as an enclosure for gas or liquid and when a hole could result in loss of fluid. Metal penetration can also be expressed in terms of a pitting factor, which is the ratio of the deepest metal penetration to the average metal penetration (determined from mass loss): (Eq 1) A pitting factor of 1 represents uniform corrosion. The larger the number, the greater the pit depth penetration. The pitting factor cannot be used when pitting or general corrosion is slight, because values of 0 or infinity are obtained in such situations, respectively. The application of statistics to the analysis of corrosion data is covered in detail in ASTM G 16 (Ref 15). The subject is discussed briefly in Ref 13 to show how statistics can be used in the evaluation of pitting data. The probability that pitting will initiate on a metal surface depends on a number of factors, including the pitting tendency of the metal, the corrosivity of the solution, the specimen area, and the duration of exposure. A pitting probability test can be used to determine the susceptibility of metals to pitting. However, this test will not provide information about the rate of propagation, and the results are applicable only to the conditions of exposure. Pitting probability P is expressed as a percentage and requires exposure of a number of specimens to a particular set of conditions (Ref 16, 17): (Eq 2) where Np is the number of specimens that pit, and N is the total number of specimens. The relationship between pit depth and area or time of exposure may vary with such factors as the environment and the metal exposed. Equations 3 and 4 are examples of the relationships that have been found to apply under certain exposure conditions. Equation 3 is the relationship between maximum pit depth D and the area A of a pipeline exposed to soil (Ref 18, 19, 20): D = bAa

(Eq 3)

where a and b are constants (each > 0) derived from the slope and the y-intercept of a straight line obtained when the logarithm of the maximum or mean pit depth for successively increasing exposed areas on the pipeline are plotted against the logarithm of the corresponding pipeline areas. The dependence on exposed area is attributed to the increased chance for the deepest pit to be found when the size of the sample of pits is increased by an increase in the area of corroded surface. There is a common but highly mistaken tendency to divide measured pit depth by exposure time to compute a pitting rate, for example, mils per year, millimeters per year, and so on. Unless independently corroborated by good data, this approach can be misleading and potentially hazardous in predicting time to failure by pitting corrosion. The maximum pit depth D of aluminum exposed to various waters was found to vary as the cube root of time t, as shown in Eq 4 (Ref 16, 21): D = Kt1/3

(Eq 4)

where K is a constant that is dependent on the composition of the water and the alloy. Equation 4 has been found to apply to several aluminum alloys exposed to different waters. Extreme value probability statistics (Ref 22, 23) have been successfully applied to estimate the maximum pit depth on a large area of material based on examination of only a small portion of that area (Ref 14, 16, 21). The procedure consists of measuring maximum pit depths on several replicate specimens and then arranging the pit depth values in order of increasing rank. A plotting position for each order of ranking is obtained by substitution in the relation M/(n + 1), where M is the order of ranking of the specimen in question, and n is the total number of specimens or values. For example, the plotting position for the second value out of 10 would be 2/(10 + 1) = 0.1818. These values are plotted on the ordinate of extreme value probability paper versus corresponding maximum pit depths. A straight line indicates that extreme value statistics are applicable.

Extrapolation of the straight line can be used to determine the probability that a specific pit depth will occur or the number of observations that must be made to find a particular pit depth. Loss in Mechanical Properties. If pitting is the predominant form of corrosion and if the density of pitting is relatively high, the change in a mechanical property relative to unexposed control specimens can be used to characterize degree of pitting. The typical properties considered for this purpose are tensile strength, elongation, fatigue strength, impact resistance, and burst pressure (Ref 24, 25). The precautions that must be taken in the application of these mechanical test procedures are covered in most standard methods. However, it must be stressed that the exposed and unexposed specimens should replicate each other as closely as possible. Therefore, consideration should be given to such factors as specimen size, edge effects, direction of rolling, and surface finish. Representative specimens of the metal are exposed to the same conditions except for the corrosive environment. The mechanical properties of the exposed and unexposed specimens are measured after the exposure, and the difference between the two results is attributed to corrosion damage. Some of these methods are better suited to the evaluation of other forms of localized corrosion, such as intergranular or stress corrosion. Therefore, their limitations must be considered. The often erratic nature of pitting and the location of pits on the specimen can affect results. In some cases, the change in mechanical properties due to pitting may be too small to provide meaningful results. Perhaps one of the most difficult problems is to separate the effects due to pitting from those caused by other forms of corrosion.

References cited in this section 13. “Standard Recommended Practice for Examination and Evaluation of Pitting Corrosion,” G 46, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Testing and Materials, 2002 14. F.A. Champion, Corrosion Testing Procedures, 2nd ed., John Wiley & Sons, 1965, p 197 15. “Standard Recommended Practice for Applying Statistics to Analysis of Corrosion Data,” G 16, Annual Book of ASTM Standards, American Society for Testing and Materials, 2002 16. B.R. Pathak, Testing in Fresh Waters, Handbook on Corrosion Testing and Evaluation, W.H. Ailor, Ed., John Wiley & Sons, 1971, p 553 17. P.M. Aziz and H.P. Godard, Influence of Specimen Area on the Pitting Probability of Aluminum, J. Electrochem. Soc., Vol 102, Oct 1955, p 577 18. G.N. Scott, Adjustment of Soil Corrosion Pit Depth Measurements for Size of Sample, Proceedings of the American Petroleum Institute, Vol 14, Section IV, American Petroleum Institute, 1934, p 204 19. M. Romanoff, Underground Corrosion, National Bureau of Standards Circular 579, U.S. Government Printing Office, 1957, p 71 20. I.A. Denison, Soil Exposure Tests, The Corrosion Handbook, H.H. Uhlig, Ed., John Wiley & Sons, 1948, p 1048 21. H.P. Godard, The Corrosion Behavior of Aluminum in Natural Waters, Can J. Chem. Eng., Vol 38, Oct 1960, p 1671 22. E.J. Gumbel, Statistical Theory of Extreme Values and Some Practical Applications, Applied Mathematics Series 33, U.S. Department of Commerce, 1954 23. P.M. Aziz, Application of the Statistical Theory of Extreme Values to the Analysis of Maximum Pit Depth Data for Aluminum, Corrosion, Vol 12, Oct 1956, p 495

24. T.J. Summerson, M.J. Pryor, D.S. Keir, and R.J. Hogan, Pit Depth Measurements as a Means of Evaluating the Corrosion Resistance of Aluminum in Sea Water, Metals, STP 196, American Society for Testing and Materials, 1957, p 157 25. R. Baboian, “Corrosion Resistant High- Strength Clad Metal System for Hydraulic Brake Line Tubing,” SAE Preprint 740290, Society for Automotive Engineers, 1972


Acknowledgment This article has been adapted from D.O. Sprowls, Evaluation of Pitting Corrosion, Corrosion, Vol 13, ASM Handbook (formerly Metals Handbook, 9th ed.), ASM International, 1987, p 231–233.


Acknowledgment This article has been adapted from D.O. Sprowls, Evaluation of Pitting Corrosion, Corrosion, Vol 13, ASM Handbook (formerly Metals Handbook, 9th ed.), ASM International, 1987, p 231–233.



A.H. Ailor, Ed., Handbook of Corrosion Testing and Evaluation, John Wiley & Sons, 1971 R. Baboian, Ed., Manual 20 Corrosion Tests and Standards: Application and Interpretation, ASTM International, 1995 M. Fontana, Corrosion Engineering, 3rd ed., McGraw-Hill, 1986

• •

H. P. Godard, W. B. Jepson, M. R. Bothwell, and R. L. Kane, The Corrosion of Light Metals, John Wiley & Sons, 1967 Z. Szklarska-Smialowska, Pitting Corrosion of Metals, NACE International, 1986

R.M. Kain, Evaluating Crevice Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 549–561

Evaluating Crevice Corrosion R.M. Kain, Consultant

Introduction CREVICE CORROSION is a form of localized corrosion that affects many alloys that normally exhibit passive behavior. Stainless steels, particularly those with little or no molybdenum, are especially prone to crevice corrosion. In neutral and acidic chloride-containing environments, some higher-alloyed stainless-type materials and related nickel-base alloys may also be susceptible to crevice attack. Although other materials, such as aluminum alloys, copper alloys, and titanium alloys, may also be susceptible to crevice-related corrosion, testing and test development have primarily focused on assessing the behavior of the stainless-type alloys. Much of this attention has been directed toward identification and development of more resistant alloys for marine applications and in certain process industries, such as pulp and paper. Over the years, diverse crevice corrosion testing methodologies have evolved and continue to find significant utility in simple immersion studies, electrochemical testing, and the development of mathematical modeling. This article discusses a number of the more frequently used crevice corrosion testing and evaluation procedures. It is recognized, however, that other techniques may exist and that their merits should be examined, especially in the context of the intended application.

R.M. Kain, Evaluating Crevice Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 549–561 Evaluating Crevice Corrosion R.M. Kain, Consultant

Aspects of Crevice Corrosion Testing Because crevices are real space dimensions between mating materials, use of the term artificial crevice test is discouraged in favor of descriptors that identify the crevice former as naturally occurring (for example, a barnacle) or man-made (for example, a washer). Man-made crevice formers can be either a metal or nonmetal and recreate the geometry of some fabricated crevices. Man-made crevice formers may or may not reproduce all conditions found in service, particularly those that are naturally occurring. Crevice corrosion tests are based on two accepted phases of crevice corrosion: initiation and propagation. Initiation refers to the breakdown of passivity and propagation or penetration as the actual occurrence of crevice corrosion. Under some conditions, breakdown may be followed by repassivation without any outward evidence of attack. For a given alloy-environment combination, the occurrence of crevice corrosion is highly dependent on the crevice geometry. Tighter crevices are more conducive to promoting initiation, but this factor may affect

the propagation behavior of materials in different ways (Ref 1). Environmental factors may affect the rate of crevice corrosion propagation. Although natural and synthetic seawater, for example, will both initiate crevice corrosion on type 316 (Unified Numbering System, or UNS, S31600) and many other stainless alloys, propagation in the two environments is often quite different (Ref 2). The overall size of the specimen in relation to the crevice (cathode-to-anode) surface area can also affect propagation behavior. Regardless of the purpose of crevice corrosion testing, the investigator should be aware that these and many other factors can affect test results. Depending on availability, natural or actual process environments may be used in laboratory immersion tests. This provides a convenient opportunity to study, for example, the effects of temperature, crevice geometry, and other factors such as surface finish. Crevice corrosion tests can be conducted at marine corrosion test sites and at laboratories associated with chemical-processing facilities.

References cited in this section 1. T.S. Lee et al., Mathematical Modelling of Crevice Corrosion of Stainless Steels, Corrosion and Corrosion Protection Proceedings, Vol 81-8, The Electrochemical Society, 1981, p 213–224 2. R.M. Kain and T.S. Lee, Recent Developments in Test Methods for Investigating Crevice Corrosion, Laboratory Corrosion Tests and Standards, STP 866, American Society for Testing and Materials, 1985, p 299–323


Guidelines for Crevice Corrosion Testing Laboratory crevice corrosion screening tests play a valuable role in the ultimate use of stainless steels in corrosive environments. Laboratory immersion tests and electrochemical studies are important tools for investigating the effects of alloy content, surface preparation, and environmental variables. However, test results should be used judiciously so as not to limit materials that may otherwise not meet contrived conditions selected to accelerate test results. ASTM G 78 provides guidance for conducting crevice corrosion tests on stainless steels and related nickel-base alloys in seawater and other chloride-containing environments (Ref 3). While its origins grew out of attempts to standardize the use of multiple- crevice assemblies, described later in this article, this guide does not promote any particular crevice former but advises the reader to be aware that crevice corrosion test results can be affected by a number of interrelated factors. In particular, crevice geometry and specimen preparation are very important aspects of testing. This standard provides guidance in the conduct of tests for flat and cylindrical specimens as well as those having irregular surfaces, such as weldments. It also reviews evaluation and reporting considerations. Importantly, Ref 3 points out that the occurrence or absence of crevice corrosion in a given test is no assurance that it will not occur under other test conditions. This is especially true if the design fails to consider those factors that may affect initiation and/or propagation behavior.

Reference cited in this section

3. “Standard Guide for Crevice Corrosion Testing of Iron-Base and Nickel-Base Stainless Alloys in Seawater and Other Chloride Containing Aqueous Environments,” G 78, Annual Book of ASTM Standards, American Society for Testing and Materials


Guidelines for Crevice Corrosion Testing Laboratory crevice corrosion screening tests play a valuable role in the ultimate use of stainless steels in corrosive environments. Laboratory immersion tests and electrochemical studies are important tools for investigating the effects of alloy content, surface preparation, and environmental variables. However, test results should be used judiciously so as not to limit materials that may otherwise not meet contrived conditions selected to accelerate test results. ASTM G 78 provides guidance for conducting crevice corrosion tests on stainless steels and related nickel-base alloys in seawater and other chloride-containing environments (Ref 3). While its origins grew out of attempts to standardize the use of multiple- crevice assemblies, described later in this article, this guide does not promote any particular crevice former but advises the reader to be aware that crevice corrosion test results can be affected by a number of interrelated factors. In particular, crevice geometry and specimen preparation are very important aspects of testing. This standard provides guidance in the conduct of tests for flat and cylindrical specimens as well as those having irregular surfaces, such as weldments. It also reviews evaluation and reporting considerations. Importantly, Ref 3 points out that the occurrence or absence of crevice corrosion in a given test is no assurance that it will not occur under other test conditions. This is especially true if the design fails to consider those factors that may affect initiation and/or propagation behavior.

Reference cited in this section 3. “Standard Guide for Crevice Corrosion Testing of Iron-Base and Nickel-Base Stainless Alloys in Seawater and Other Chloride Containing Aqueous Environments,” G 78, Annual Book of ASTM Standards, American Society for Testing and Materials


Specific Crevice Corrosion Tests In-Plant and Field Testing. A considerable amount of crevice corrosion data has been obtained from in-plant and field test rack exposures of candidate alloys in various process streams and other industrial environments.

ASTM G 4 (Ref 4) provides guidance on specimen preparation, test rack assembly, and exposure considerations. Simple tetrafluoroethylene (TFE) fluorocarbon sleeves and washers, intended primarily as electrical insulating spacers between candidate alloy specimens, have proved to be effective crevice formers. Likewise, metal-to- metal configurations have also been used. Test results for square, rectangular, and circular spool disk specimens have identified the occurrence of crevice corrosion and the extent of penetration incurred during the exposure period. Although these and other tests lack definition with regard to actual times to initiation and rates of propagation, they are nonetheless quite effective in screening materials for further consideration. Some test racks are spring loaded, while others are not. It is generally recognized that rapid and severe attack on some specimens in a test assembly may influence crevice tightness on others in the same assembly. One area in which test rack exposures have been extensively used is the pulp and paper industry. For example, Ref 5 discusses results from bleach plant exposure in which crevice corrosion susceptibility was identified for 24 of the 26 alloys tested. However, the crevices found in bleach plant equipment may be less severe than those on the test specimens. This is supported by the observation that some materials that exhibited less than 0.4 to 0.46 mm (15 to 18 mils) of crevice penetration in a 3 month test actually performed satisfactorily in service for up to 20 years. Ferric Chloride Crevice Corrosion Tests. Ferric chloride tests involve exposure to a highly oxidizing acid chloride environment. The pH of the standard 6% FeCl3 solution, according to ASTM G 48 (Ref 6), is approximately 1.2. This is less than the pH of the critical crevice solution, which can cause a breakdown of passivity for UNS S30400 and S31600 stainless steels. Critical crevice solution compositions for these alloys with respect to pH and chloride concentration have been identified as pH 2.0 at 1.5 M Cl- and pH 1.7 at 3.5 M Cl-, respectively (Ref 7, 8). Use of FeCl3 eliminates the time factor for development of the aggressive crevice solution from the neutral pH of, for example, a chloride- containing natural water. The crevice prevents oxygen from maintaining the passive film. The combination of low pH and high Cl- lowers the breakdown or critical potential for pitting in the crevice, while ferric ions (Fe3+) maintain the potential of the alloy in the critical region that supports the cathodic reaction outside the crevice. More resistant alloys do not exhibit a breakdown potential in the domain associated with Fe3+ reduction at ambient temperature. In some cases, the Fe3+ serves as a chemical potentiostat and holds the corrosion potential well within the passive region in the absence of oxygen. Increasing the FeCl3 solution temperature to 50 °C (120 °F), for example, promotes breakdown and allows for ranking of more resistant alloys. Figure 1 shows schematically the polarization characteristics for stainless steels exposed to FeCl3.

Fig. 1 Schematic showing the polarization characteristics of stainless-type materials in an acid chloride solution with and without Fe3+. Ecorr, corrosion potential; i, current density ASTM and Materials Technology Institute (MTI) Test Methods. Both ASTM G 48 (method B) and MTI-2 (Ref 9) used a simple 6% ferric chloride solution prepared with FeCl3·6H2O and distilled water solutions. Both are intended for short-term testing within ambient and elevated- temperature ranges. In the G 48 (method B) procedure, crevices are formed by the placement of a polytetrafluoroethylene (PTFE) cylindrical block on each side of the test specimen and securing them in place with rubber bands. Tests at 22 ± 2 °C (72 ± 3.5 °F) are conducted for 72 h. Crevices in the MTI-2 method are created by the application of two modified multiplecrevice assemblies (serrated PTFE washers), each having 12 plateaus or contact sites. An assembly torque of 0.28 N · m (2.5 in. · lbf) was prescribed for tests of 24 h duration. The MTI-2 method calls for a series of tests so as to define a critical crevice temperature (CCT) within 2.5 °C (4.5 °F) increments. A CCT was one that produced at least a 0.025 mm (0.001 in.) deep attack at one or more of the available crevice sites. There are currently two ASTM methods available for CCT testing: one intended primarily for nickel-base and chromium-containing alloys, G 48 (method D), and one for stainless steels, G 48 (method F). Corresponding standardized pitting tests appear under designations G 48 (methods A, C, and E). Both G 48 (methods D and F) use the serrated washer device mentioned previously. Method D specifies the original MTI-2 torque of 0.28 N · m (2.5 in. · lbf), while method E specifies a higher torque of 1.58 N · m (14 in. · lbf). Currently, the recommended test durations are 12 and 24 h, respectively. Both methods D and F use an acidified (1% HCl) solution of ferric chloride (6% FeCl2), intended to provide stable pH conditions during elevated-temperature testing. During modification of the original MTI-2 method, ASTM subcommittee G-01.05 adopted two alloychemistry-based equations useful in identifying a starting test temperature for the respective methods. Both G 48 (methods D and F) are intended to identify a CCT within the 0 to 85 °C (32 to 185 °F) range by the use of

fresh specimens for each exposure. If crevice corrosion is observed at the initial test temperature, methods D and F advise lowering the temperature for the next test by 5 °C (9 °F) and repeating the test. If no attack is found after the initial tests, the temperature is raised by 10 °C (18 °F) and a new specimen exposed. The process is repeated, as needed, until the CCT is identified. The criterion for crevice corrosion in these methods is a measured attack depth of ≥0.025 mm (0.001 in.). Precision and bias statements exist for ASTM G 48 (methods D and F). For the purpose of alloy ranking, specimens tested by all three methods (B, D, and F) can be evaluated by mass loss, number of sites attacked, and the range of penetrations incurred. The user is cautioned that surface preparation can influence test results. Testing in ferric chloride solution with fresh specimens for each temperature increment resulted in lower CCT values than when the same specimen was reused without resurfacing (Ref 10). ASTM G 48 has been used for alloy development purposes as well as for quality control. Generally, acceptable criteria are established by agreement between alloy producer and the purchaser. The value of CCT ranking for service life prediction remains limited. Some correlation has been reported for CCT and specific seawater test data (Ref 11, 12), but it does not take into account the effect of crevice geometry and the interrelationship with the natural environment. Although increasing temperature (for example, from 25 to 50 °C, or 77 to 120 °F) of the FeCl3 solution is certainly detrimental, the same is not necessarily true with respect to crevice corrosion in seawater where bulk oxygen concentration decreases with increasing temperature (due to inverse solubility), and biological activity may increase initially but decreases drastically above a threshold temperature (Ref 13). Sodium Chloride Crevice Corrosion Test. Materials Technology Institute method MTI-4 (Ref 9) involves testing in chloride solutions at eight concentrations, ranging from 0.1 to 3.0% NaCl, to establish a minimum (critical) Cl- concentration that results in crevice corrosion at room temperature (20 to 24 °C, or 68 to 75 °F). Specimen requirements, crevice assembly procedure, and evaluation technique for the MTI-4 method are identical to those used in the MTI-2 method. The recommended test period for the MTI-4 method, however, is 1000 h, and suggested controls are UNS S30400 and S31600 stainless steel. Both methods are acceptable screening tools and are well suited for alloy development purposes. However, extension of either method beyond its intended scope may be misleading. In the case of MTI-4, for example, deviations in crevice geometry from that of the serrated washer could significantly alter resistance to a given bulk Cl- concentration. Tighter and/or larger crevices could promote initiation at lower Cl- levels (Ref 14). Use of the serrated washer identified earlier, for example, has produced crevice initiation at Cl- levels as low as 100 to 300 mg/L in a matter of a few days when assembled at an initial torque of 8.5 N · m (75 in. · lbf) (Ref 15). Metal-to-metal crevice sites can sometimes promote more rapid initiation and at lower Cl- levels in comparison to nonmetal-to- metal crevices of comparable geometry (Ref 16).

References cited in this section 4. “Standard Guide for Conducting Corrosion Tests in Field Application,” G 4, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Testing and Materials 5. A.H. Tuthill, Resistance of Highly Alloyed Materials and Titanium to Localized Corrosion in Bleach Plant Environments, Mater. Perform., Vol 24 (No. 9), 1985, p 43–49 6. “Standard Method for Pitting and Crevice Corrosion Resistance of Stainless Steels and Related Alloys by the Use of Ferric Chloride Solution,” G 48, Annual Book of ASTM Standards, American Society for Testing and Materials 7. J.W. Oldfield and W.H. Sutton, Crevice Corrosion of Stainless Steels, Part II: Experimental Studies, Br. Corros. J., Vol 13 (No. 3), 1978, p 104 8. R.M. Kain et al., Use of Electrochemical Techniques for the Study of Crevice Corrosion in Natural Seawater, Section 6: Localized Corrosion, Electrochemical Techniques for Corrosion Engineering, NACE Publication, 1986, p 261–279

9. R.S. Treseder and E.A. Kachik, MTI Corrosion Tests for Iron and Nickel-Base Corrosion Alloys, Laboratory Corrosion Tests and Standards, STP 866, American Society for Testing and Materials, 1985, p 373–399 10. E.L. Hibner, Modification of Critical Crevice Temperature Test Procedures for Nickel Alloys in a Ferric Chloride Environment, Mater. Perform., Vol 26 (No. 3), 1987, p 37–40 11. A.P. Bond and H.J. Dundas, Resistance of Stainless Steels to Crevice Corrosion in Seawater, Mater. Perform., Vol 23 (No. 7), 1984, p 39–43 12. A. Garner, Crevice Corrosion of Stainless Steel in Seawater: Correlation of Field Data with Laboratory Ferric Chloride Tests, Corrosion, Vol 37 (No. 3), 1981, p 178–184 13. T.S. Lee et al., The Effect of Environmental Variables on Crevice Corrosion of Stainless Steels in Seawater, Mater. Perform., Vol 23 (No. 7), 1984, p 9–15 14. R.M. Kain, Crevice Corrosion Behavior of Stainless Steel in Seawater and Related Environments, Corrosion, Vol 40 (No. 6), 1984, p 313–321 15. R.M. Kain et al., The Resistance of Type 304 and Type 316 Stainless Steel to Crevice Corrosion Natural Waters, J. Mater. Energy Syst., Vol 5 (No. 4), 1984, p 205–211 16. J.W. Oldfield and W.H. Sutton, New Technique for Predicting the Performance of Stainless Steels in Seawater and Other Chloride Containing Environments, Br. Corros. J., Vol 15 (No. 1), 1980, p 31–34


Multiple-Crevice Assembly Testing The multiple-crevice assembly (MCA) technique was originally introduced in the mid- 1970s as a rapid and economical screening tool for establishing resistance to crevice corrosion in natural seawater (Ref 17). Testing was specifically intended to be severe enough to produce some indication of alloy susceptibility or resistance instead of relying on other factors, such as fouling and seasonal variation in temperature. It was also desirable that the MCA demonstrate recognized differences in alloy capabilities, such as those between UNS S30400 and S31600 stainless steels, as well as the exceptional degree of performance expected of highly corrosion-resistant alloys. Unlike some other popular crevice corrosion tests, the MCA test relied solely on naturally occurring processes and required neither outside electrochemical nor chemical stimulation. It was argued that because crevice corrosion appeared to be random in its occurrence, the creation of a number of identical crevice sites on a given set of specimens would provide a basis for statistical analysis. Therefore, the MCA design and the use of the probability-of-attack concept gained appeal (Ref 18). Early concerns with geometrical factors primarily addressed the size of the specimen in terms of boldly exposed area (cathode) to shielded or crevice area (anode) (Ref 17). Subsequent research, especially through mathematical modeling (Ref 1, 14, 16), showed that the occurrence of crevice corrosion was dependent on a number of interrelated factors, while the probability concept suggested only a material (i.e., alloy) property. Because crevice geometry factors, such as crevice tightness, could be overriding considerations (Ref 19, 20), proponents and users of the MCA

generally abandoned the probability concept in describing the occurrence of crevice corrosion. However, as evidenced by the number of investigations reported in the literature (for example, see Ref 21, 22, 23, 24), the MCA technique remained popular, because it enabled the investigator to report results in terms of initiation and propagation. For example, resistance to initiation could be expressed as the number of crevice sites or the percentage of sites exhibiting attack in a given test period. Because attack, when it occurs, is located at specific sites (Fig. 2), penetration measurements could be made at each site and reported as a maximum depth. This can be accomplished by using an appropriate dial depth gage or focusing microscope. Results have been reported as a range of these maximum values or their average value. Table 1, for example, gives data describing the effect of molybdenum content on both initiation and propagation resistance for a series of controlled-chemistry heats of a nickel- base alloy containing 30% Cr. Multiple-crevice test results have been reported in bar graph form as mean values plus and minus one standard deviation of the percentage of sites initiated and maximum depth of attack (Ref 25).

Fig. 2 Example of crevice attack associated with 20 plateau multiple-crevice assembly. Courtesy of the LaQue Center for Corrosion Technology, Inc. Table 1 Multiple-crevice assembly test results showing the beneficial effects of molybdenum in nickelbase alloys containing 30% Cr Results of 30 day test in filtered, natural seawater at 30 °C (85 °F) Alloy Alloy molybdenum Initiation resistance Penetration resistance content, % Time observed, h Average number of sites Maximum Average visibly corroded depth depth Earliest Average mm mils mm mils 0 52 76 10–11 1.75 69 0.8 31.5 A 0.5 52 76 11 1.65 65 0.48 19 B 0.9 100 115 9–10 1.2 47 0.4 16 C 2.6 100 115 8–9 0.8 31.5 0.2 8 D 4.9 ND ND 1 0.05 2 0.05 2 E ND, not detected during 30 day test. Source: Ref 21 Reference 26, which was prepared under a grant from the National Institute of Standards and Technology (NIST), provides a critical review on the significance and use of MCAs for testing in natural seawater. The report traces the development and changes in design and application of the acetal resin device shown in Fig. 2. Included in Ref 26 are results of an ASTM International subcommittee G-01.09 round-robin test aimed at standardization of the MCA test procedure. Triplicate specimens of four different alloys, with wet-ground (120grit SiC) surfaces and MCAs tightened to an initial torque of 8.5 N · m (75 in. · lbf), were exposed for 30 days

in natural seawater. Of the eight participants with access to natural seawater, seven were located on the U.S. east coast at latitudes ranging from Key West, FL, to New England. The eighth test location was the California coast. All except two test sites provided some degree of seawater filtration; three were unable to provide temperature control. As expected, some differences in seawater chemistry (for example, chloride and dissolved oxygen) and ambient temperatures were reported. Although the purpose of these tests was to determine the reproducibility of results for different classes of materials, the test alloys were also ranked for corrosion susceptibility. None of the participants reported any incidence of crevice corrosion for nickel alloy C-276 (UNS N10276), which contains ~15% Mo. On the other hand, at least one specimen of alloy G (UNS N06007), which contains ~6% Mo, initiated attack at all eight test locations. Overall, the actual number of crevice sites attacked and the range and maximum depth of penetration for UNS N06007 were both found to be quite minimal. For UNS S30400 and S31600 stainless steels, however, crevice corrosion occurred on all specimens except for one UNS S31600. In general, identification of alloy susceptibility was quite reproducible. However, in terms of the actual number of sites that initiated crevice corrosion, much variability was observed, particularly with respect to test location and participant. Because of the variability in the round-robin test results, a second set of tests with five participants was conducted at one location. The test was restricted to UNS S30400 and S31600 stainless steels. Actual assembly and conduct of the test were performed by individuals with different degrees of experience in MCA testing. All of the tests were conducted at the same time in natural seawater at 25 ± 2 °C (77 ± 3.5 °F). The results of these tests are shown in Fig. 3. Overall, the degree of variability is consistent with that observed from data provided from the original round-robin testing at different test locations. In addition, penetrations measured for the two stainless steels in the latter round-robin test were comparable to those in the original.

Fig. 3 Multiple-crevice assembly round-robin results from 30 day test conducted simultaneously at one ASTM International selected seawater site. AISI, American Iron and Steel Institute

In another series of 30 day comparative tests (Ref 27), 13 participants used MCAs in round- robin testing of UNS S30400 in 3.5% NaCl at 30 °C (85 °F). Crevice assembly design and initial torque were the same as in the ASTM International committee G-01.09 round-robin test. Again, considerable variability was evident from one participant to another. Crevice sites that exhibited initiation out of 120 available crevice sites ranged from 20 to 86%. Although some exceptions were noted, a number of individual participants reported good reproducibility among replicate specimens. Compared to attack in natural seawater, penetrations in 3.5% NaCl tests did not exceed 0.23 mm (9 mils). Average penetrations for triplicate specimens varied by only a few hundredths of a millimeter. The degree of variability described previously is likely to be encountered in any type of crevice corrosion test. When conducted in a conscientious manner, MCA tests can generate meaningful results for ranking of alloys and identifying the effects of other exposure variables. The MCA test cannot and should not be used to predict alloy performance under other conditions. Another test series involved exposure of several grades of stainless steel in different to pulp- bleaching environments. Testing was performed in accordance with a procedure developed by NACE International T-5H6 task group (“Test Methods for Measuring Crevice Corrosion Rates”). Multiple exposures involved the use of a Rulon (Saint-Gobain Performance Plastics) crevice device similar to that shown in Fig. 2 but with somewhat larger plateaus. Participants used specimens of a given size, and the assemblies were torqued to prescribed levels. Because only a few crevice sites initiated, and reproducibility was poor, it was concluded that the method was not useful in warm (up to 82 °C, or 180 °F), low- pH (1 to 5), low-chloride (50 to 2000 ppm) environments. However, these conditions actually fit well with the premise that a range of tight crevice gaps exists for each type of assembly. Mathematical modeling has shown an interrelationship between bulk chloride level and crevice gap (Ref 14). For the relatively small depth of a MCA site and the low-level chlorides present in the bleach plant, only the tightest of sites should be expected to initiate crevice corrosion.

References cited in this section 1. T.S. Lee et al., Mathematical Modelling of Crevice Corrosion of Stainless Steels, Corrosion and Corrosion Protection Proceedings, Vol 81-8, The Electrochemical Society, 1981, p 213–224 14. R.M. Kain, Crevice Corrosion Behavior of Stainless Steel in Seawater and Related Environments, Corrosion, Vol 40 (No. 6), 1984, p 313–321 16. J.W. Oldfield and W.H. Sutton, New Technique for Predicting the Performance of Stainless Steels in Seawater and Other Chloride Containing Environments, Br. Corros. J., Vol 15 (No. 1), 1980, p 31–34 17. D.B. Anderson, Statistical Aspects of Crevice Corrosion in Seawater, Galvanic and Pitting Corrosion— Field and Laboratory Studies, STP 576, American Society for Testing and Materials, 1976, p 261 18. A.P. Bond et al., Corrosion Resistance of Stainless Steels in Seawater, Advanced Materials in Seawater Applications, Climax Molybdenum Company, Piacenza, Italy, Feb 1980, p 1 19. J.W. Oldfield, “Crevice Corrosion of Stainless Steels: The Importance of Crevice Geometry and Alloys Composition,” presented at 19th Journees de Aciers Speciaux, (Saint Etienne, France), May 1980, Met. Corros.- Ind., April 1981, No. 668, p 134–147 20. T. Sydberger, Werkst. Korros., Vol 32 (No. 3), 1981, p 119 21. R.M. Kain, “Effect of Alloy Content on the Localized Corrosion Resistance of Several Nickel Base Alloys in Seawater,” Paper 229, presented at Corrosion/86, National Association of Corrosion Engineers, 1986 22. R.M. Kain, “Crevice Corrosion and Metal Ion Concentration Cell Corrosion Resistance of Candidate Materials for OTEC Heat Exchangers,” ANL/OTEC-BCM-022, Argonne National Laboratory and the U.S. Department of Energy, May 1981

23. G.O. Davis and M.A. Streicher, “Initiation of Chloride Crevice Corrosion on Stainless Alloys,” Paper 205, presented at Corrosion/ 85, National Association of Corrosion Engineers, 1985 24. H.P. Hack, Crevice Corrosion Behavior of 45 Molybdenum Containing Stainless Steels in Seawater, Mater. Perform., Vol 22 (No. 6), 1983, p 24–30 25. T.S. Lee and R.M. Kain, “Factors Influencing the Crevice Corrosion Behavior of Stainless Steels in Seawater,” Paper 69, presented at Corrosion/83, National Association of Corrosion Engineers, 1983 26. R.M. Kain, Crevice Corrosion Testing in Natural Seawater: Significance and Use of Multiple Crevice Assemblies, J. Test. Eval., Sept 1990, p 309–319 27. U. Hideki et al., “Crevice Corrosion of Stainless Steels in Chloride Solutions,” Paper 117, presented at Corrosion/89, National Association of Corrosion Engineers, 1989


Continuous Crevice Formers Perspex Crevice Assemblies. While MCAs were being used extensively in the 1980s to investigate crevice corrosion in marine and paper mill environments, another type of device was developed as a research tool. The Perspex crevice assembly (PCA) was initially devised and used to provide immersion test verification data for the development of mathematical models of crevice corrosion (Ref 7). Perspex is a British trade name (Imperial Chemical Industries, UK) for a hard plastic, polymethyl methacrylate (PMMA). This material and related plastics, such as Plexiglas (ATOFINA Chemicals, Inc.), can be machined as a continuous-type crevice former with depth dimensions that are intentionally varied. The design shown in Fig. 4 was prepared with bored holes that enabled access of the test medium to the inside diameter (ID), creating an effective crevice depth equivalent to one-half the PCA annular dimension. The most common- size PCAs reported in the literature are those measuring 40 mm outside diameter (OD) by 20 mm ID (1.6 by 0.8 in.) and a smaller version measuring 30 mm OD by 20 mm ID (1.2 by 0.8 in.) (Ref 3, 28). Because of the translucent nature of the acrylic plastic, the earliest stages of crevice corrosion development can be detected in situ. This led to greater accuracy in reporting times to initiation and the extent of measured propagation.

Fig. 4 Perspex crevice assembly with ports to allow aqueous environment access to inside diameter. Courtesy of the LaQue Center for Corrosion Technology, Inc. Perspex crevice assemblies have been used extensively in comparing alloy performance and in tests to investigate variations in Cl- and levels in simulated natural waters, crevice geometry, test temperature, alloy surface conditions, and cathode-to-anode area ratio (Ref 28, 29, 30, 31, 32). Perspex crevice assemblies have been used in the smooth, polished, ground, and machine- grooved condition to vary the severity of the geometry formed at a particular assembly torque, usually 2.8 N · m (25 in. · lbf). In some tests, crevice conditions were made even more severe by introducing a deformable insert (plastic tape) between the hard plastic and the test metal surface. Effects of the previously mentioned variations on alloy performance in natural seawater are reported in Ref 31 and 32. Results for several alloys that have diverse chemistries are shown in Table 2. Table 2 Identification of metallurgical/ geometrical test conditions that promoted crevice corrosion of various alloys in 60 day natural seawater Perspex crevice assembly test Test material Alloy resistance (R) or susceptibility (S) for indicated test conditions(a) (b) UNS No. CRN A B C D E F G H 24 S S S S S S S S S31603 37 R R R R R R S S S32550 43 R R R R R R S R S31254 43 R S R R R R S S Alloy VI 45 R R R R R R R S N08926 45 R R R R R R R R N08367 R R R R R R S S Alloy VIII 46 51 R S R R R R R R N08031 56 R R R R R R R R S32654 69 R R R R R R S R N10276 73 R R R R R R R R N06059 UNS, Unified Numbering System. (a) A, specimen surface, as-produced mill; Perspex crevice assembly (PCA) washer finish, 0.05 μm alumina polish. B, specimen surface, as-produced mill; PCA washer finish, A and tape insert. C, specimen surface, as-

produced mill; PCA washer finish, 600-grit SiC. D, specimen surface, as-produced mill; PCA washer finish, 120-grit SiC. E, specimen surface, as-produced mill; PCA washer finish, 60-grit SiC. F, specimen surface, asproduced mill; PCA washer finish, lathe turned (Ra, 15 μm). G, specimen surface, ground (120-grit SiC); PCA washer finish, 0.05 μm alumina polish. H, specimen surface, ground and pickled (10 min in 15% HNO3 + 6% HF at 40 °C, or 105 °F); PCA washer finish, 0.05 μm alumina polish. (b) CRN, corrosion resistance number (%Cr + 3.3 × %Mo + 16 × %N), equal to pitting resistance equivalent number, PREN. Source: Ref 32 For the most part, PCAs have been assembled with simple nut, bolt, and washer assemblies of a corrosionresistant material such as UNS N10276 or titanium Gr-2 (UNS R50400). As described, for example, in Ref 19, precision washers have also been used to ensure a constant pressure. At least one investigation (Ref 33) has employed continuous washers of the 30 mm OD by 20 mm ID (1.2 by 0.8 in.) design but fashioned from other polymers, including PTFE, Rulon, and fiber-impregnated PTFE. In these tests, the crevice formers were made without access holes and thus had a nominal crevice depth of 5.0 mm (0.2 in.) as contrasted with the 2.5 mm (0.1 in.) depth inherent with the original PCA. Unlike the translucent, polished acrylic plastic, PCAs made from the previously mentioned materials do not permit detection of the early stages of crevice corrosion initiation. Gaskets. Over the years, gaskets have been a perennial source of crevice corrosion at flanged surfaces and other mounting sites where a nonmetallic was used for sealing or for electrical isolation purposes. In early seawater tests, for example, crevice corrosion often developed beneath Bakelite (Georgia Pacific Corp.) insulators used to mount specimens on test racks. As reported in Ref 34, PTFE and cloth fiber- filled rubber gaskets were used in an extensive U.S. Navy program to evaluate candidate valve materials for seawater service. Materials considered for valve body applications were tested with grooved surfaces and the fiber-filled rubber gaskets. Other material considered for valve trim hardware was tested in the surface-ground condition with PTFE gaskets. In both tests, the respective metal surface roughness was quantified. Both types of gasket were ~50 mm OD by 13 mm ID (2 by 0.5 in.) and held in place on both sides of the 100 mm by 150 mm (4 by 6 in.) testpieces, as shown in Fig. 5. Tables 3 and 4 summarize the results of a 6 month test conducted in filtered seawater at 25 ± 5 °C (77 ± 9 °F).

Fig. 5 Exploded (a) and completed (b) views of gasket crevice assembly of the type used in Ref 34 and elsewhere. Gasket held in place with polyvinyl chloride and titanium retainer. Courtesy of the LaQue Center for Corrosion Technology, Inc. Table 3 Summary of results for wrought stainless steel and nickel-base alloys in 6 month natural seawater tests with polytetrafluoroethylene (PTFE) gaskets

UNS No. or other designation Number of sites attacked (4 max) Maximum depth of attack mm in. 4 2.27 0.089 S31603 3 1.17 0.046 N08367 4 1.73 0.068 S31254 0 0.0 0.000 S32654 4 1.04 0.041 S39275 4 0.98 0.039 S32760 2 0.59 0.023 S21000 3 0.18 0.007 N06625 2 0.13 0.005 N10276 0 0.00 0.000 N06022 0 0.00 0.000 N06686 0 0.00 0.000 N06059 0 0.00 0.000 C-2000 4 0.43 0.017 N07716 3 1.18 0.046 N07718 4 3.18 0.125 N09925 Note: PTFE gasket site machined to 0.5 to 1.5 μm finish. UNS, Unified Numbering System. Source: Ref 34 Table 4 Summary of results for various cast alloys in 6 month natural seawater test with clothimpregnated rubber gaskets UNS No. or other designation Number of sites attacked (4 max) Maximum depth of attack mm in. 2 0.34 0.013 C92200 4 0.60 0.024 C95800 4 0.71 0.028 N24135 4 0.25 0.010 C96400 4 3.75 0.148 J92800 (a) 4 3.97 0.156 N08007 4 3.39 0.133 J94651 (a) 4 3.31 0.130 J93254 2 1.66 0.065 N26625 3 2.83 0.111 N30002 1 0.98 0.039 N26022 0 0.00 0.000 Alloy 59 cast Note: Except for cast alloy 59, all other materials were thin-section investment castings and tested in as-cast condition. UNS, Unified Numbering System. (a) UNS N08007 and J93254 were tested with smooth surfaces (Table 3), while all others were tested with a coarse phonographic finish. Source: Ref 34 The effect of gasket material on the crevice corrosion resistance of cast (UNS J92800) and wrought (UNS S31600) stainless steel was investigated in seawater and a less aggressive Cl-- and -containing water (Ref 35). Identification of 11 gasket materials (plus PCA controls) and seawater test results for J92800 flanges are provided in Table 5. In seawater, the propensity for attack favored crevices formed with nitrile and PTFE-type gaskets. The nominal cathode-to- crevice area ratio in the tests was smaller than in many of the other crevice corrosion tests cited here. Table 5 Effect of selected gasket material on the crevice corrosion initiation and propagation behavior of cast alloy UNS J92800 (CF3M) flanges in seawater Gasket material

Number of flanges attacked Maximum depth of attack

mm in. 2.10 0.083 Amaride fiber + nitrile 2/2 1/2 1.40 0.055 Glass-filled PTFE 2/2 1.05 0.041 PTFE 0.77 0.030 Carbon fiber + nitrile 2/2 0.69 0.027 Graphite/stainless steel 1/2 1/2 0.05 0.002 EPDM 1/2 0.03 0.001 Nitrile 2/2 0.01 0.000 PCA-type control 1/2 0.01 0.000 Red rubber 0/2 0.00 0.000 Fluoroelastomer 0/2 0.00 0.000 Butyl rubber 0/2 0.00 0.000 Neoprene UNS, Unified Numbering System; PTFE, polytetrafluoroethylene; EPDM, ethylene-propylene-diene monomer (rubber); PCA, Perspex crevice assembly. Note: Flanges surface ground (Ra, 0.4 μm). Testing comprised 21 days exposure to flowing (0.7 m/s, or 2.3 ft/s) seawater, followed by 7 days wet lay-up, resulting in stagnant conditions. Source: Ref 35 Irregular Surfaces. The use of rigid crevice formers with deformable inserts, for example, PCAs and gasket assemblies, can generally produce tight crevices on metals with irregular surfaces associated with coarse grinding, machining, and heavy acid pickling. However, these types of devices are generally not applicable to testing autogenous welds, raised weld beads, or unfinished weld overlays. Some success has been achieved by using O-rings or rods (Ref 36) made of deformable polymers. In each case, some type of holder or retainer is needed to position and secure the crevice former over the desired section of weldment. This may be achieved by bolting through a hole drilled in the test specimen or by using a clamping device. Another, more adaptable method involves the use of an epoxy-type coating. Reference 37 describes the application of an epoxy coating to assess the crevice corrosion behavior of weld metal and adjacent heataffected zones on cruciform-shaped specimens in seawater. Reference 37 discusses the repeatability aspects of the testing performed on UNS S31603, S20910, and N08367. Elsewhere, a paint marking pen was successfully used to create crevice conditions and assess the performance of seal-welded tube- to-tube sheet assemblies fabricated with unidentified stainless steel components (Ref 38). Figure 6 shows before-and-after views of a tested section. The use of dried plastic from a felt tip pen to produce crevice sites is discussed in Ref 39. Some of the attributes described for this technique include a constant critical geometry, no need for fixturing, a crevice area that can be varied, and simple specimens having concave, convex, or uneven surfaces.

Fig. 6 Example of the use of paint marker (a) to create crevice site on irregular weld surface and (b) resulting attack after exposure to seawater. Courtesy of the LaQue Center for Corrosion Technology, Inc.

There appears to be some similarity in the undulated pattern of attack found beneath coatings (see Fig. 6b) and that associated with heat tint on as-welded stainless steel exposed to seawater and other chloride-containing solutions.

References cited in this section 3. “Standard Guide for Crevice Corrosion Testing of Iron-Base and Nickel-Base Stainless Alloys in Seawater and Other Chloride Containing Aqueous Environments,” G 78, Annual Book of ASTM Standards, American Society for Testing and Materials 7. J.W. Oldfield and W.H. Sutton, Crevice Corrosion of Stainless Steels, Part II: Experimental Studies, Br. Corros. J., Vol 13 (No. 3), 1978, p 104 19. J.W. Oldfield, “Crevice Corrosion of Stainless Steels: The Importance of Crevice Geometry and Alloys Composition,” presented at 19th Journees de Aciers Speciaux, (Saint Etienne, France), May 1980, Met. Corros.- Ind., April 1981, No. 668, p 134–147 28. J.W. Oldfield et al., Avoiding Crevice Corrosion of Stainless Steels, Proc. Stainless Steel '84 Symposium, (Gotenberg, Sweden), Chalmers University of Technology and Jernkontoret (Sweden) and The Metals Society (UK), 1984 29. R.M. Kain and J.W. Oldfield, “Prediction of Crevice Corrosion Resistance of Stainless Steels in Aqueous Environments—A Corrosion Engineering Guide,” presented at 12th International Corrosion Congress, Localized Corrosion Session, (Houston, TX), NACE International, 1993 30. J.W. Oldfield and R.M. Kain, Assessment of the Corrosion Resistance of Austenitic Stainless Steels in Industrial Waters, Proc. International Corrosion Congress, (Florence, Italy), Associazione Italiana Di Metallurgia, 1990 31. R.M. Kain, “Effects of Surface Finish on the Crevice Corrosion Resistance of Stainless Steels in Seawater and Related Environments,” Paper 508, presented at Corrosion/ 91, National Association of Corrosion Engineers, 1991 32. R.M. Kain, Seawater Testing to Assess the Crevice Corrosion Resistance of Stainless Steels and Related Alloys, Proc. 12th International Corrosion Congress, Localized Corrosion Session, Vol 3, (Houston, TX), 1993, NACE International, p 1889–1900 33. P.A. Klein et al., “The Effect of Electrolytic Chlorination on the Crevice Corrosion Behavior of 70/30 Copper-Nickel and Nickel- Copper Alloy 400,” Paper 509, presented at Corrosion/91, National Association of Corrosion Engineers, 1991 34. D.M. Aylor et al., “Crevice Corrosion Performance of Candidate Naval Ship Seawater Valve Materials in Quiescent and Flowing Natural Seawater,” Paper 329, presented at Corrosion/99, NACE International, 1999 35. R.M. Kain, Gasket Materials and Other Factors Influencing the Crevice Corrosion Resistance of Stainless Steel Flanges, Mater. Perform., Vol 37 (No. 8), 1998, p 62–66 36. R.A. Buchanan and C.D. Lundin, New Method for Crevice Corrosion Testing of Welds with Reinforcements, Corrosion, Vol 58 (No. 5), 2002, p 448 37. R.M. Kain, Use of Coatings to Assess the Crevice Corrosion Resistance of Stainless Steels in Warm Seawater, Marine Corrosion in Tropical Environments, STP 1399, American Society for Testing and Materials, 2000

38. LaQue Center for Corrosion Technology, Inc., Unpublished test report, 2000 39. T. Degerbeck and T. Gille, Crevice Corrosion—A New Crevice Former, Corros. Sci., Vol 19, 1979, p 1113–1114


Evaluating Cylindrical Materials and Products It is often desirable to evaluate the crevice corrosion resistance of product forms other than flat sheet and plate. Crevice corrosion testing of rod and barstock as well as alloy tubing and pipe can be accomplished by the use of devices that can produce a crevice on the OD of the test material. Such devices exist as off-the-shelf fittings or other products that can be cut and adapted to meet test needs. Among those available are wood collars, O-rings, ferrule and sleeve-type compression fittings, polyvinyl chloride (PVC) compression fittings with internal rubber glands, various flexible hose products, snap-type hose clamps, packing materials, and epoxy coatings. The selection of an appropriate crevice former should take into account the purpose of the tests, for example, direct one-on-one alloy evaluations, assessing environmental variables, or evaluating product performance under conditions likely to be found in service. The selection depends heavily on the size (OD) of the product being tested and the available size of the crevice- forming material. The fact that rod and barstock can be machined to specific dimensions gives the investigator some control over the test design. In contrast, the testing of tubular products may be dictated by the available mill-produced OD and whether the products are formed to English or metric dimensions. The combination of product form and selected crevice device can accurately reproduce or closely approximate the crevice geometries found in some service applications. O-Rings. The simplest method of crevice corrosion testing of cylindrical specimens, including tubulars, is by the application of O-rings (PTFE, rubber, Viton [DuPont Dow Elastomers] or other polymer). The crevice can be made more or less severe by varying the size of the O-ring with respect to the OD of the alloy product. It should be recognized that excessive stretching of an O-ring may alter its width and, accordingly, the depth of the crevice formed. O-rings of different thicknesses allow the crevice depth to be varied. One advantage of Orings is that they do not require tightening and thus eliminate a source of variation found in some other types of devices. However, O-rings should not be overstretched. Another advantage of O-rings is that results from simple immersion tests can be correlated with electrochemical tests, such as ASTM F 764 (Ref 40), that use a rod-shaped electrode with an O-ring crevice former. A disadvantage is that the crevice depth is limited. The shallow depth of an O-ring is less likely to promote attack of more highly alloyed stainless steels and nickel-base alloys that otherwise may be more susceptible with deeper crevices, for example, with gasket or washer-type crevice formers. As a rule, deep- tight crevices are considered to be more severe than shallow-loose crevices. Wood Collars. Collars fashioned from water- resistant redwood (but not copper-chromium- arsenate-treated lumber) have been used to assess the relative resistance of stainless alloy shaft stock to crevice corrosion in seawater (Ref 41). In this case, the redwood was drilled to provide holes of the same diameter as the OD of the barstock tested. The lumber was then split to provide two half-bracelets and further cut into the shape of collars analogous to a penal stock. The collars were positioned over a length of barstock material and tightly secured with corrosion-resistant fasteners. The crevice depth can be intentionally changed by varying the lumber thickness. The cathode-to-anode (crevice) area ratio can be varied by varying the bar length or by adding multiple collars.

Packing Material. Packing glands are sites where crevice corrosion could occur, for example, on boat and pump shafts. Reference 42 details a method for comparing the effect of different types of packing materials on the crevice corrosion resistance of stainless steel shafting under static and dynamic conditions. In brief, the packing material was cut into lengths approximately one-half the circumference of the test material. The lengths of packing were freely shaped to form a collar or gland around the test-piece and held tightly in place within a retainer fashioned from a PVC fitting. Some internal machining of the fitting may be required to ensure a tight fit of the packing around the testpiece. Sleeve Assemblies. The use of sleeves (Fig. 7) produced from polymer and rubber tubing or from hose and applied to the OD of cylindrical specimens allows crevices of varying tightness and depth to be formed. Under simple immersion conditions, these nonmetallic sleeves can be used alone, without the need for additional clamping. For direct alloy comparisons, the OD of all test specimens should be the same, so that crevice tightness is not a variable.

Fig. 7 Translucent vinyl sleeve assembly applied to pipe or tube outside diameter for crevice corrosion testing under static or dynamic flow conditions. Courtesy of the LaQue Center for Corrosion Technology, Inc. The vinyl sleeve test method offers several advantages. These include ease of preparation, ability to produce tight crevices, and the translucent nature, which facilitates detection of crevice corrosion initiation at the earliest stages. A number of investigations (Ref 43, 44, 45, 46, 47, 48) have used vinyl sleeves (Tygon tubing, Saint-Gobain Performance Plastics) and/or rubber hose to test alloy tube and pipe sections under dynamic conditions. In those tests, seawater (treated and untreated) flowed through a series of alloy tubes or pipes connected by the vinyl sleeves. In dynamic tests, it is essential that the sleeves be clamped to the specimens to prevent separation due to line pressure. The spacing between specimens should be minimal to prevent any bulging of the sleeve that could vary crevice geometry. In tests of this design, the wetted interior pipe surface (ID) serves as the cathode and the shielded portion of the pipe OD as the crevice (anode). Ideally, all tubular specimens should be of the same size and schedule. However, this is not always possible. Reference 45 describes one seawater crevice corrosion test program that examined the performance of a series

of high-alloy stainless steel produced to English and metric dimensions. In these tests, the IDs of the vinyl sleeves either matched or were smaller than the ODs of the tubular test specimens. Testing proceeded with the understanding that the severity of the crevices was different. In Table 6, those specimens tested with the tightest fit were identified as severe (S) and those with matching ID/ ODs as moderately severe (MS). The differences in seawater velocity associated with each specimen ID dimension are also noted in Table 6. Table 6 Crevice corrosion resistance exhibited by 10 different stainless-type tubing alloys in 60 day flowing seawater/vinyl sleeve tests UNS No.

Seawater velocity m/s ft/s

Crevice severity(a)

Initiation resistance Earliest days 4 4 7 14 14 14 43 53 OK at 60 OK at 60

detection, Number of sites attacked(b) (6 max) 6 6 6 3 3 3 4 2 0 0

2.4 7.9 MS S31603 2.4 7.9 MS S31703 4.0 13.1 MS S31254 2.4 7.9 MS S32750 10.8 S N08925 3.3 6.6 MS N08367 2.0 2.4 7.9 MS S32550 5.2 S N08926 1.6 2.4 7.9 MS S32760 2.4 7.9 MS S39724 UNS, Unified Numbering System. (a) MS, moderately severe; S, severe. (b) Maximum depth (0.57 mm, or 0.022 in.) found on UNS S31703 tube; all others ≤0.22 mm (0.0087 in.). Source: Ref 45 Large-Diameter Pipe. Testing the crevice corrosion resistance of large-diameter pipe (e.g., 250 mm, or 10 in.) creates challenges for the investigator. In a previously unreported case (Ref 49), large-diameter pipes prepared by different metallurgical processes were tested in seawater in lengths measuring approximately 750 mm (30 in.). Equally spaced holes were drilled along the length of the pipe at 0°, 90°, 180°, and 270°. To accommodate the inner and outer curvature of the pipes, sections of 25.4 mm diameter by 6.4 mm (1 by 0.25 in.) wall vinyl tubing were split and punched into 25.4 mm (1 in.) OD diameter disks with a 6.4 mm (0.25 in.) center hole. To create the desired crevice on the OD and ID of the test pipe, these disks were then mated with the ID and OD of the pipe at the location of the drilled holes and bolted in place with micarta retainers. The resulting conditions produced crevice corrosion of alloy UNS N06625 within 60 days and revealed no significant effect of the metallurgical processes that were investigated. Compression Fittings. A disadvantage of vinyl sleeves is their cost relative to that of other crevice formers, particularly when testing larger, for example, >25.4 mm (1 in.), alloy tubulars. One option is to use inexpensive PVC compression fittings with internal rubber glands. The width of the gland increases with the pipe size of the fitting. Some variability in crevice geometry may result during assembly and tightening of the compression nuts on either end. Selection of this type of crevice-forming device, while economical, sacrifices the ability for early in situ detection of crevice attack available with the translucent vinyl sleeve device. However, if the attack is significant, corrosion products may become evident at the compression nut/tube interface. References 50 and 51 cite results of two seawater test programs aimed at examining benefits of surface treatment to enhance the crevice corrosion resistance of stainless steel and nickel-base alloy UNS N06625. In another test program (Ref 52), the effects of alternating flow and stagnant seawater conditions on the behavior of 90/10 copper-nickel (UNS C71500) and alloy- 400 (UNS N04400) were investigated with this type of crevice former. Elsewhere, the behavior of stainless alloy tubing fitted with PVC compression fittings was investigated in aerated, deaerated, and sulfide-containing seawater (Ref 52). Other types of readily available but more costly compression fittings containing internal ferrules and sleeves have also been used. One test program (Ref 53) used UNS S31600 stainless steel compression fittings to test the resistance of two versions of a 6Mo-containing stainless steel in a tube-to-tubesheet simulation. Nylon compression fittings of this type have also been used to compare alloy resistance to natural and chlorinated

seawater (Ref 43). Results were compared with those from concurrent tests with vinyl sleeves and those made from Buna-N (E.I. Dupont) tubing. The ferrule-and-sleeve devices actually create at least three different crevice geometries within an assembly: a shallow-tight one beneath the ferrule, a deeper-tight crevice at the sleeve, and a considerably deeper but less tight crevice between the tube wall and fitting annulus. Depending on alloy sensitivity, attack may occur at all locations or only at one or both of the tighter crevice sites. Due to their opaqueness and need for tightening, the ferrule-and-sleeve-type devices share the same limitations as the PVC/rubber gland devices. On the other hand, some versions are available in both English and metric unit dimensions and are therefore suitable for testing tubular and other cylindrical products. Metal-to-Metal Assemblies. Reference 52 describes metal-to-metal crevice corrosion testing of stainless steel tubulars fitted with compression fittings. A test simulating a tube-to-tube support sheet (Fig. 8) configuration with 0.13 mm (0.005 in.) clearance was reported in Ref 45. While both tubes and plate comprising UNS S30400 stainless steel were attacked within 60 days, no attack of AL-6XN (UNS N08367) tubes or mating tube support plates of UNS S31603 or alloy 2507 (UNS S37250) occurred.

Fig. 8 Example of mixed-metal crevice assembly intended to mimic actual tube and tube support plate conditions (described in Ref 45). Courtesy of the LaQue Center for Corrosion Technology, Inc. Coatings, such as a marine-grade epoxy, have been used to create crevice conditions on both cylindrical and flat specimens (Ref 37, 54, 55). This is a relatively inexpensive approach for testing of tubulars of all sizes under simple immersion conditions. The coating can be applied to as-produced and prepared tubular surfaces (Ref 54). Anode-to-cathode area ratio effects can be investigated by varying the percentage of tube surface area coated (Fig. 9). In general, the interface between the coating (edge) and the bare metal serves as the crevice mouth and from which corrosion product inevitably appears if the material is susceptible. Once initiated, the attack may continue further inward beneath the coating, progressively disbonding it.

Fig. 9 Use of epoxy coating to create crevice conditions on alloy pipe specimen. Coated surface varied to investigate area ratio effect. Courtesy of the LaQue Center for Corrosion Technology, Inc. Assessing Corrosion Damage on Cylindrical Specimens. As with flat-panel-type specimens, cylindrical and tubular specimens can be evaluated by their resistance to initiation and propagation. Initial resistance can be reported according to the observed times to visible attack as well as the number or percent of crevice sites attacked. Depending on the size of the specimens, it may be possible to determine mass loss. In addition, affected crevice area and the depth of attack (penetration) can be determined. Affected area can be measured (or estimated) by placing a transparent grid over the crevice site and counting the number of area units (e.g., millimeters squared). Measuring the depth of attack for a cylindrical specimen can be more challenging than for a flat specimen. One method entails placing the cylindrical specimen in a lathe and rotating the specimen manually. A dial depth gage mounted on the traversing tool post is then used to measure depth of attack along the circumference and axis of the specimen. For example, depth of attack can be measured along a line at various degrees of rotation, for example, every 15°, 30°, or 45°.

References cited in this section 37. R.M. Kain, Use of Coatings to Assess the Crevice Corrosion Resistance of Stainless Steels in Warm Seawater, Marine Corrosion in Tropical Environments, STP 1399, American Society for Testing and Materials, 2000 40. “Standard Test Method for Pitting or Crevice Corrosion of Metallic Surgical Implant Materials,” F 746, Annual Book of ASTM Standards, Vol 13.01, American Society for Testing and Materials 41. LaQue Center for Corrosion Technology, Inc., Unpublished test report, 1978

42. R.M. Kain and I. Dunoff, “Influence of Packing Material on the Corrosion Resistance of Stainless Steel Boat Shafting and Related Materials,” Paper 639, presented at Corrosion/2000, NACE International, 2000 43. R.M. Kain et al., “Crevice Corrosion of Nickel-Chromium-Molybdenum Alloys in Natural and Chlorinated Seawater,” Paper 112, presented at Corrosion/89, National Association of Corrosion Engineers, 1989 44. R.M. Kain and P.A. Klein, “Crevice Corrosion Propagation Studies for Alloy N06625: Remote Crevice Assembly Testing in Flowing Natural and Chlorinated Seawater,” Paper 158, presented at Corrosion/90, National Association of Corrosion Engineering, 1990 45. R.M. Kain and A. Zeuthen, Crevice Corrosion Testing of Austenitic, Superaustenitic, Superferritic and Superduplex Stainless Type Alloys in Seawater, Corrosion Testing in Natural Waters, STP 1300, American Society for Testing and Materials, 1996 46. M.E. Inman et al., “Detection of Crevice Corrosion in Natural Seawater Using Polarization Resistance Measurements,” Paper 297, presented at Corrosion/97, NACE International, 1997 47. D.C. Agarwal, “Solving Critical Problems in Marine Environments by an Advanced Ni-Cr-Mo Alloy 59 UNS N06059,” Paper 635, presented at Corrosion/2000, NACE International, 2000 48. R.M. Kain, “Testing for Crevice Corrosion Susceptibility,” presented at research symposium at Corrosion/96, NACE International, 1996 49. LaQue Center for Corrosion Technology, Inc., Unpublished test report, 1993 50. R.M. Kain and M.B. Ives, “Surface Treatments Benefiting the Crevice Corrosion Resistance of Nickel Alloy N06625 in Natural and Chlorinated Seawater,” Paper 330, presented at Corrosion/99, NACE International, 1999 51. R.M. Kain et al., “Localized and General Corrosion Resistance of Candidate Metallic Materials for ROMembrane Cartridges,” Paper 268, presented at Corrosion/95, NACE International, 1995 52. R.M. Kain, A. Zeuthen, and J. Maurer, “Localized Corrosion Resistance of Stainless Type Materials in Aerated, Deaerated, and Stagnant Sulfide Bearing Seawater,” Paper 423, presented at Corrosion/97, NACE International, 1997 53. P.A. Klein et al., “A Localized Corrosion Assessment of 6% Molybdenum Stainless Steel Condenser Tubing at the Calvert Cliffs Nuclear Power Plant,” Paper 490, presented at Corrosion/94, National Association of Corrosion Engineers, 1994 54. R.M. Kain, “Crevice Corrosion Behavior of Coated Stainless Steel in Natural Seawater,” Paper 827, presented at Corrosion/2000, NACE International, 2000 55. R.M. Kain, “Seawater Crevice Corrosion Resistance of Stainless Steels Coated with Silane and Antifouling Paint Systems,” Paper 187, presented at Corrosion/2002, NACE International, 2002


Component Testing In some cases, it may be possible to test manufactured components or assemblies and assess their resistance to crevice corrosion. These may comprise nonmetal-to-metal or metal-to-metal joints. In the latter case, like-metal or dissimilar- metal couples may be involved. Testing of flanges with gaskets or O-rings and fasteners are common examples. In an unpublished test (Ref 56), a number of highly torqued like-metal and dissimilar-metal nut, bolt, and washer assemblies were tested by alternate immersion in seawater for 25 weeks. Fasteners comprising all UNS N06625 components exhibited some evidence of crevice attack at mated nut-to-washer, bolthead-to-washer, washer-to-washer, nut-to- bolthead and nut-to-bolt-thread surfaces. When NiCu alloy UNS N04400 washers were substituted for UNS N06625, no attack on the latter occurred, due to galvanic protection (sacrificial corrosion) provided by the NiCu alloy. In another case using UNS N06625 washers in an otherwise all-titanium assembly, attack of the nickel-base alloy was limited to the mated washer surfaces. No attack was found at the titanium-to-UNS N06625 crevice sites created at bolt-head and nut locations. In a metal-to-metal configuration, hydrolysis reactions involving corrosion products from both surfaces in a crevice can contribute to changes in the crevice electrolyte chemistry, leading to breakdown of passivity (Ref 7). Only the UNS N06625 surface in the preceding example would be a contributor of pH-lowering chromium ions. Reference 57 describes seawater testing of alloy 686 (UNS N08686) bolting securely fastened to large plates of UNS N06625. Specimens were exposed with and without cathodic protection from zinc anodes.

References cited in this section 7. J.W. Oldfield and W.H. Sutton, Crevice Corrosion of Stainless Steels, Part II: Experimental Studies, Br. Corros. J., Vol 13 (No. 3), 1978, p 104 56. LaQue Center for Corrosion Technology, Inc., Unpublished test report, 1998 57. E.L. Hibner and L.E. Shoemaker, “High Strength Corrosion Resistance Alloy 686 for Seawater Fastener Service,” Paper 195, presented at Corrosion/2002, NACE International, 2002


Electrochemical Tests ASTM Standard Methods. ASTM G 61 is recommended primarily for investigating localized corrosion on iron, nickel-, or cobalt-base alloys (Ref 58). Crevices are formed on a 16 mm (0.63 in.) diameter disk test electrode by a TFE- fluorocarbon gasket/mounting assembly. Other gasket materials can also be used. The test electrode

is the anodic member of a polarization cell containing a deaerated 3.5% NaCl solution. After a 1 h period of free corrosion, the test specimen electrode potential is scanned in the noble direction at a rate of 0.6 V/h. Current density and potential are continuously plotted. On reaching an anodic current density of 5 mA/cm2, the scan direction is reversed and the scan continued back to its original potential. Susceptibility to crevice corrosion is identified by the occurrence of positive hysteresis during the reverse scan. Relative alloy resistance can be established by comparing the forward and reverse scan potential-current domains with UNS N10276 and S30400 stainless steel standards. ASTM G 61 provides standard polarization plots for comparison and equipment performance verification. Some degree of variability is expected. Several factors may affect results, most notably the actual time between specimen preparation and exposure as well as the degree of tightness used in assembling the test electrode holder. Nonstandard variations of ASTM G 61, such as the use of different electrode assemblies and/ or polarization scan rates, may significantly affect the measured response (Ref 59). Some correlations have been made between cyclic polarization tests performed according to ASTM G 61 and immersion crevice tests in seawater (Ref 60). For the most part, cyclic polarization tests are able to differentiate between highly alloyed, resistant materials (for example, the high-molybdenum nickel-base alloys UNS N06625 and N10276) and lower-alloy materials (for example, 300-series stainless steel). Correlation for alloys with intermediate compositions may be more difficult, because they exhibit resistance to a broader range of environments (that is, Cl- concentration and crevice geometry) than the 300- series stainless steels but less than that for nickel-base chromium plus highmolybdenum alloys. Previously mentioned ASTM F 746 describes procedures for evaluating the pitting and crevice corrosion resistance of metallic surgical implant materials (Ref 40). Testing is again limited to making relative rankings of performance. The procedure calls for the use of a cylindrical electrode with crevices formed by a mounting compression gasket and an intentional crevice-forming collar. Corrosion is induced by a polarization step to +0.8 V versus a saturated calomel reference electrode (SCE). Provisions are established for current monitoring and subsequent potential-time increments. The test is designed to produce corrosion for a control electrode of UNS S31600 stainless steel. Alloys can be compared relative to breakdown, propagation, and repassivation tendencies. Other Potentiostatic and Potentiodynamic Polarization Tests. A potentiostatic test procedure (Ref 61) identifies alloy crevice corrosion resistance according to an established critical crevice temperature (CCT). Tests have been conducted in neutral NaCl solution and synthetic seawater under constant applied potentials, for example, +600 mV versus SCE. In this automated test, the equipment is programmed to increase the solution temperature in 5 °C (9 °F) increments until a specific critical current level is reached in a given period, for example, 15 to 20 min. Test results obtained with the previously mentioned method using MCAs are summarized in Table 7 (Ref 62). Also given in Table 7 is a CCT ranking determined in 72 h FeCl3 tests performed according to ASTM G 48 (rubber band test) and another ranking based on the total number of crevice corrosion sites initiated in 30 day natural seawater MCA tests. The three procedures gave the same ranking merit for only 3 of the 12 test materials, numbers 1, 2, and 6. In several cases, two procedures provided the same ranking for a given alloy. In the Santron test, the noble potential is intended to mimic the redox potential of FeCl3. In natural seawater, such potentials would never be achieved without chemical stimulation. In the absence of crevice corrosion, nickelbase chromium-molybdenum alloys as well as some stainless steels have achieved potentials in the +300 to +400 mV range versus SCE in ambient temperature seawater (Ref 55, 63, 64). Table 7 Initiation of crevice corrosion in immersion tests in seawater, FeCl3, and synthetic seawater All specimens were ground with 120-grit SiC. Alloy Filtered seawater(a)

29-4C Monit SC-1

Number sites 0 0 1

FeCl3(b) Failure temperature of Rank °C °F 1 2 3

55 47 45

131 117 113

Synthetic seawater(c) Rank Failure temperature °C °F

Rank

1 2 4

1 2 5

90.0 67.5 60.0

195 155 140

2 4 37 99 5 60.0 140 4 Ferralium 255 5 28 82 8 47.5 120 7 Haynes No. 20 6 Mod 11 6 37 99 6 57.5 135 6 AL6X 18 7 46 115 3 62.5 145 3 254SMO 36 8 22 72 10 42.5 110 9 904L 47 9 31 88 7 45.0 115 8 JS700 60 10 14 57 12 30.0 85 12 JS777 73 11 25 77 9 40.0 105 11 AISI type 329 112 12 15 59 11 40.0 105 10 Nitronic 50 (a) 30 day test at 30 °C (85 °F). (b) 72 h test in 10% FeCl3·6H2O. (c) Santron test; 20 min measuring time. Source: Ref 62 Potentiodynamic polarization tests for crevice-free electrodes in a series of increasingly aggressive simulated crevice solutions have been used to rank alloys according to a criterion associated with anodic peak current density (Ref 7). Such information has been used in mathematical modeling to identify the localized environment or critical crevice solution (CCS) that can cause breakdown of passivity and the conditions leading up to it (Ref 28, 30). In addition, as shown in Fig. 10, a plot of log current versus pH has been used to characterize propagation resistance for several cast alloys (Ref 65). Alloy propagation behavior can be ranked according to the slope of the log i/pH plot. For the previously mentioned tests, the electrode and its wire lead connection are potted with a thermosetting resin. After curing, the electrode surface is ground, then wiped with a coat of resin to fill gaps, and reground or polished prior to testing.

Fig. 10 Plot of anodic peak current density versus simulated crevice solution pH used to determine the composition (pH) of the critical crevice solution (CCS) according to the 10 μA/cm2 (64.5 μA/in.2) criterion. Source: Ref 8 While not intended for the previously mentioned purpose, ASTM G 150 (Ref 66) may also have some utility in CCS identification. ASTM G 150 was developed to produce a crevice-free electrode for the study of critical pitting temperature. While a crevice gap is present between the test electrode and its holder, this space is continuously flushed with purified water to prevent the buildup of aggressive H+ and Cl- ions that would otherwise contribute to breakdown of passivity and lead to crevice attack. Remote Crevice Assemblies. An electrochemical procedure requiring no stimulation other than that provided by the bulk environment/alloy reaction has been described in the literature. Techniques have been identified as either remote crevice or remote cathode tests (Ref 44, 65, 67, 68). Remote crevice assembly tests involve the physical separation but electrical connection of a small anode or crevice member and a larger cathodic member. Both are exposed in the bulk environment. Current between the two members can be monitored through a zeroresistance ammeter. The technique is quite capable of accurately identifying the time to initiation and subsequent propagation.

Unlike other techniques, this procedure is able to separate the initiation and propagation phases of crevice corrosion. This capability is summarized by the plot of corrosion current normalized for initiation time in Fig. 11. Results show good reproducibility in both the trend of increasing current once initiation has occurred as well as the magnitude of current. The total charge (coulombs) is a measure of propagation, which is directly proportional to mass loss due to crevice corrosion.

Fig. 11 Comparison of crevice corrosion propagation currents for UNS S31603 stainless steel remote crevice assemblies after normalizing initiation times. Source: Ref 2 Other Electrochemical Techniques. Other specialized tests have been developed, with varying degrees of success. One test involves the use of compartmentalized cells with anode and cathode members exposed to actual or simulated environments (Ref 69). Such tests have been used to evaluate the effects of changes in solution chemistry and surface area ratios. In addition, a vibrating electrode technique has been used to map variations in current density above a creviced stainless steel specimen of known crevice geometry (Ref 70). Such tests are more expensive and perhaps more conducive to mechanistic studies rather than for routinely characterizing alloy resistance to crevice corrosion. A number of electrochemical techniques that may be considered for crevice corrosion testing are reviewed in Ref 71 and 72.

References cited in this section 2. R.M. Kain and T.S. Lee, Recent Developments in Test Methods for Investigating Crevice Corrosion, Laboratory Corrosion Tests and Standards, STP 866, American Society for Testing and Materials, 1985, p 299–323 7. J.W. Oldfield and W.H. Sutton, Crevice Corrosion of Stainless Steels, Part II: Experimental Studies, Br. Corros. J., Vol 13 (No. 3), 1978, p 104

8. R.M. Kain et al., Use of Electrochemical Techniques for the Study of Crevice Corrosion in Natural Seawater, Section 6: Localized Corrosion, Electrochemical Techniques for Corrosion Engineering, NACE Publication, 1986, p 261–279 28. J.W. Oldfield et al., Avoiding Crevice Corrosion of Stainless Steels, Proc. Stainless Steel '84 Symposium, (Gotenberg, Sweden), Chalmers University of Technology and Jernkontoret (Sweden) and The Metals Society (UK), 1984 30. J.W. Oldfield and R.M. Kain, Assessment of the Corrosion Resistance of Austenitic Stainless Steels in Industrial Waters, Proc. International Corrosion Congress, (Florence, Italy), Associazione Italiana Di Metallurgia, 1990 40. “Standard Test Method for Pitting or Crevice Corrosion of Metallic Surgical Implant Materials,” F 746, Annual Book of ASTM Standards, Vol 13.01, American Society for Testing and Materials 44. R.M. Kain and P.A. Klein, “Crevice Corrosion Propagation Studies for Alloy N06625: Remote Crevice Assembly Testing in Flowing Natural and Chlorinated Seawater,” Paper 158, presented at Corrosion/90, National Association of Corrosion Engineering, 1990 55. R.M. Kain, “Seawater Crevice Corrosion Resistance of Stainless Steels Coated with Silane and Antifouling Paint Systems,” Paper 187, presented at Corrosion/2002, NACE International, 2002 58. “Standard Practice for Conducting Cyclic Potentiodynamic Polarization Measurements for Localized Corrosion,” G 61, Annual Book of ASTM Standards, Vol 03.02, American Society for Testing and Materials 59. R.M. Kain, “Localized Corrosion Behavior in Natural Seawater: A Comparison of Electrochemical and Crevice Testing of Stainless Steel,” Paper 70, presented at Corrosion/80, National Association of Corrosion Engineers, 1980 60. B.E. Wilde, A Critical Appraisal of Some Popular Laboratory Electrochemical Tests for Predicting the Localized Corrosion Resistance of Stainless Alloys in Sea Water, Corrosion, Vol 28 (No. 8), 1972, p 283 61. S. Bernhardsson, Paper 85, presented at Corrosion/80, National Association of Corrosion Engineers, 1980 62. N.S. Nagaswami and M.A. Streicher, “Accelerated Laboratory Tests for Crevice Corrosion of Stainless Alloys,” Paper 7, presented at Corrosion/83, National Association of Corrosion Engineers, 1983 63. J.M. Kroughman and F.P. Ijsseling, Crevice Corrosion of Stainless Steels and Nickel Alloys in Seawater, Proc. Fifth International Congress on Marine Corrosion and Fouling, G. Londres, Ed., (Barcelona, Spain), 1980, p 214 64. A. Mollica et al., Cathodic Performance of Stainless Steels in Natural Seawater as a Function of Microorganism Settlement and Temperature, Corrosion, Vol 48 (No. 1), 1998, p 48–56 65. T.S. Lee et al., Use of Electrochemical Techniques for the Study of Crevice Corrosion in Natural Seawater, Electrochemical Techniques for Corrosion Engineering, National Association of Corrosion Engineering, 1986 66. “Standard Test Method for Electrochemical Critical Pitting Temperature Testing of Stainless Steels,” G 150, Wear and Erosion, Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Testing and Materials

67. T.S. Lee, A Method of Quantifying the Initiation and Propagation Stages of Crevice Corrosion, Electrochemical Corrosion Testing, STP 727, American Society for Testing and Materials, 1981, p 43– 68 68. R.M. Kain, Electrochemical Measurement of the Crevice Corrosion Propagation Resistance of Stainless Steels: Effect of Environmental Variables, Mater. Perform., Vol 23 (No. 2), 1984, p 24 69. R.M. Kain and T.S. Lee, “The Effect of Crevice Solution pH on Corrosion Behavior of Stainless Steels,” Paper 27, presented at Corrosion/84, National Association of Corrosion Engineers, 1984 70. H.S. Issacs, “Application of the Vibration Probe to Localized Current Measurements,” Paper 55, presented at Corrosion/85, National Association of Corrosion Engineers, 1985 71. J. Postlewaite, Can. Metall. Q., Vol 22 (No. 1), 1983, p 133 72. F.P. Ijsseling, Electrochemical Methods in Crevice Corrosion Testing, Br. Corros. J., Vol 15 (No. 2), 1980, p 51


Mathematical Modeling Mathematical models have been developed to serve as research and predictive tools describing the many interrelated factors known to influence crevice corrosion behavior in chloride-containing environments (Ref 7, 73, 74). Model development is heavily dependent on input from laboratory confirmation tests using some of the evaluation methods described in this article. The PCA test, for example, was used extensively in support of the Sutton-Oldfield model (Ref 7, 16) and its subsequent refinements (Ref 73). In addition, results from potentiodynamic polarization tests were used to model passive current and CCS chemistries (Ref 16). Figure 12 shows a sample model output describing the influence of crevice geometry on initiation behavior of UNS S31803 stainless steel (Ref 73). Figure 13 shows a sample model prediction of expected resistance for UNS S31600 in various chloride- and sulfate-containing waters (Ref 73).

Fig. 12 Example of Crevice Corrosion Engineering Guide (Ref 73) model output describing effect of crevice geometry on stainless steel resistance

Fig. 13 Model prediction showing effect of chloride and sulfate ion on a range of stainless steels Reference 74 describes a model that addresses both crevice corrosion initiation and propagation as well as passivation. A number of citations provided in Ref 74 and 75 offer additional reading on this subject.

References cited in this section 7. J.W. Oldfield and W.H. Sutton, Crevice Corrosion of Stainless Steels, Part II: Experimental Studies, Br. Corros. J., Vol 13 (No. 3), 1978, p 104 16. J.W. Oldfield and W.H. Sutton, New Technique for Predicting the Performance of Stainless Steels in Seawater and Other Chloride Containing Environments, Br. Corros. J., Vol 15 (No. 1), 1980, p 31–34 73. J.W. Oldfield and R.M. Kain, Prediction of Crevice Corrosion Resistance of Stainless Steels in Aqueous Environments—A Corrosion Engineering Guide, Proc. 12th International Corrosion Congress, National Association of Corrosion Engineers, 1993, p 1876 74. P.O. Gartland, Modelling Crevice Corrosion of Fe-Cr-Ni-Mo Alloys in Chloride Solutions, Proc. 12th International Corrosion Congress, Vol 3B, National Association of Corrosion Engineers, 1993, p 1901– 1914 75. B. Shaw, P. Moran, and P.O. Gartland, Crevice Corrosion of a Nickel-Based Super Alloy in Natural and Chlorinated Seawater, Proc. 12th International Corrosion Congress, Vol 3B, National Association of Corrosion Engineers, 1993, p 1915–1928


References 1. T.S. Lee et al., Mathematical Modelling of Crevice Corrosion of Stainless Steels, Corrosion and Corrosion Protection Proceedings, Vol 81-8, The Electrochemical Society, 1981, p 213–224 2. R.M. Kain and T.S. Lee, Recent Developments in Test Methods for Investigating Crevice Corrosion, Laboratory Corrosion Tests and Standards, STP 866, American Society for Testing and Materials, 1985, p 299–323 3. “Standard Guide for Crevice Corrosion Testing of Iron-Base and Nickel-Base Stainless Alloys in Seawater and Other Chloride Containing Aqueous Environments,” G 78, Annual Book of ASTM Standards, American Society for Testing and Materials 4. “Standard Guide for Conducting Corrosion Tests in Field Application,” G 4, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Testing and Materials 5. A.H. Tuthill, Resistance of Highly Alloyed Materials and Titanium to Localized Corrosion in Bleach Plant Environments, Mater. Perform., Vol 24 (No. 9), 1985, p 43–49 6. “Standard Method for Pitting and Crevice Corrosion Resistance of Stainless Steels and Related Alloys by the Use of Ferric Chloride Solution,” G 48, Annual Book of ASTM Standards, American Society for Testing and Materials

7. J.W. Oldfield and W.H. Sutton, Crevice Corrosion of Stainless Steels, Part II: Experimental Studies, Br. Corros. J., Vol 13 (No. 3), 1978, p 104 8. R.M. Kain et al., Use of Electrochemical Techniques for the Study of Crevice Corrosion in Natural Seawater, Section 6: Localized Corrosion, Electrochemical Techniques for Corrosion Engineering, NACE Publication, 1986, p 261–279 9. R.S. Treseder and E.A. Kachik, MTI Corrosion Tests for Iron and Nickel-Base Corrosion Alloys, Laboratory Corrosion Tests and Standards, STP 866, American Society for Testing and Materials, 1985, p 373–399 10. E.L. Hibner, Modification of Critical Crevice Temperature Test Procedures for Nickel Alloys in a Ferric Chloride Environment, Mater. Perform., Vol 26 (No. 3), 1987, p 37–40 11. A.P. Bond and H.J. Dundas, Resistance of Stainless Steels to Crevice Corrosion in Seawater, Mater. Perform., Vol 23 (No. 7), 1984, p 39–43 12. A. Garner, Crevice Corrosion of Stainless Steel in Seawater: Correlation of Field Data with Laboratory Ferric Chloride Tests, Corrosion, Vol 37 (No. 3), 1981, p 178–184 13. T.S. Lee et al., The Effect of Environmental Variables on Crevice Corrosion of Stainless Steels in Seawater, Mater. Perform., Vol 23 (No. 7), 1984, p 9–15 14. R.M. Kain, Crevice Corrosion Behavior of Stainless Steel in Seawater and Related Environments, Corrosion, Vol 40 (No. 6), 1984, p 313–321 15. R.M. Kain et al., The Resistance of Type 304 and Type 316 Stainless Steel to Crevice Corrosion Natural Waters, J. Mater. Energy Syst., Vol 5 (No. 4), 1984, p 205–211 16. J.W. Oldfield and W.H. Sutton, New Technique for Predicting the Performance of Stainless Steels in Seawater and Other Chloride Containing Environments, Br. Corros. J., Vol 15 (No. 1), 1980, p 31–34 17. D.B. Anderson, Statistical Aspects of Crevice Corrosion in Seawater, Galvanic and Pitting Corrosion— Field and Laboratory Studies, STP 576, American Society for Testing and Materials, 1976, p 261 18. A.P. Bond et al., Corrosion Resistance of Stainless Steels in Seawater, Advanced Materials in Seawater Applications, Climax Molybdenum Company, Piacenza, Italy, Feb 1980, p 1 19. J.W. Oldfield, “Crevice Corrosion of Stainless Steels: The Importance of Crevice Geometry and Alloys Composition,” presented at 19th Journees de Aciers Speciaux, (Saint Etienne, France), May 1980, Met. Corros.- Ind., April 1981, No. 668, p 134–147 20. T. Sydberger, Werkst. Korros., Vol 32 (No. 3), 1981, p 119 21. R.M. Kain, “Effect of Alloy Content on the Localized Corrosion Resistance of Several Nickel Base Alloys in Seawater,” Paper 229, presented at Corrosion/86, National Association of Corrosion Engineers, 1986 22. R.M. Kain, “Crevice Corrosion and Metal Ion Concentration Cell Corrosion Resistance of Candidate Materials for OTEC Heat Exchangers,” ANL/OTEC-BCM-022, Argonne National Laboratory and the U.S. Department of Energy, May 1981 23. G.O. Davis and M.A. Streicher, “Initiation of Chloride Crevice Corrosion on Stainless Alloys,” Paper 205, presented at Corrosion/ 85, National Association of Corrosion Engineers, 1985

24. H.P. Hack, Crevice Corrosion Behavior of 45 Molybdenum Containing Stainless Steels in Seawater, Mater. Perform., Vol 22 (No. 6), 1983, p 24–30 25. T.S. Lee and R.M. Kain, “Factors Influencing the Crevice Corrosion Behavior of Stainless Steels in Seawater,” Paper 69, presented at Corrosion/83, National Association of Corrosion Engineers, 1983 26. R.M. Kain, Crevice Corrosion Testing in Natural Seawater: Significance and Use of Multiple Crevice Assemblies, J. Test. Eval., Sept 1990, p 309–319 27. U. Hideki et al., “Crevice Corrosion of Stainless Steels in Chloride Solutions,” Paper 117, presented at Corrosion/89, National Association of Corrosion Engineers, 1989 28. J.W. Oldfield et al., Avoiding Crevice Corrosion of Stainless Steels, Proc. Stainless Steel '84 Symposium, (Gotenberg, Sweden), Chalmers University of Technology and Jernkontoret (Sweden) and The Metals Society (UK), 1984 29. R.M. Kain and J.W. Oldfield, “Prediction of Crevice Corrosion Resistance of Stainless Steels in Aqueous Environments—A Corrosion Engineering Guide,” presented at 12th International Corrosion Congress, Localized Corrosion Session, (Houston, TX), NACE International, 1993 30. J.W. Oldfield and R.M. Kain, Assessment of the Corrosion Resistance of Austenitic Stainless Steels in Industrial Waters, Proc. International Corrosion Congress, (Florence, Italy), Associazione Italiana Di Metallurgia, 1990 31. R.M. Kain, “Effects of Surface Finish on the Crevice Corrosion Resistance of Stainless Steels in Seawater and Related Environments,” Paper 508, presented at Corrosion/ 91, National Association of Corrosion Engineers, 1991 32. R.M. Kain, Seawater Testing to Assess the Crevice Corrosion Resistance of Stainless Steels and Related Alloys, Proc. 12th International Corrosion Congress, Localized Corrosion Session, Vol 3, (Houston, TX), 1993, NACE International, p 1889–1900 33. P.A. Klein et al., “The Effect of Electrolytic Chlorination on the Crevice Corrosion Behavior of 70/30 Copper-Nickel and Nickel- Copper Alloy 400,” Paper 509, presented at Corrosion/91, National Association of Corrosion Engineers, 1991 34. D.M. Aylor et al., “Crevice Corrosion Performance of Candidate Naval Ship Seawater Valve Materials in Quiescent and Flowing Natural Seawater,” Paper 329, presented at Corrosion/99, NACE International, 1999 35. R.M. Kain, Gasket Materials and Other Factors Influencing the Crevice Corrosion Resistance of Stainless Steel Flanges, Mater. Perform., Vol 37 (No. 8), 1998, p 62–66 36. R.A. Buchanan and C.D. Lundin, New Method for Crevice Corrosion Testing of Welds with Reinforcements, Corrosion, Vol 58 (No. 5), 2002, p 448 37. R.M. Kain, Use of Coatings to Assess the Crevice Corrosion Resistance of Stainless Steels in Warm Seawater, Marine Corrosion in Tropical Environments, STP 1399, American Society for Testing and Materials, 2000 38. LaQue Center for Corrosion Technology, Inc., Unpublished test report, 2000 39. T. Degerbeck and T. Gille, Crevice Corrosion—A New Crevice Former, Corros. Sci., Vol 19, 1979, p 1113–1114

40. “Standard Test Method for Pitting or Crevice Corrosion of Metallic Surgical Implant Materials,” F 746, Annual Book of ASTM Standards, Vol 13.01, American Society for Testing and Materials 41. LaQue Center for Corrosion Technology, Inc., Unpublished test report, 1978 42. R.M. Kain and I. Dunoff, “Influence of Packing Material on the Corrosion Resistance of Stainless Steel Boat Shafting and Related Materials,” Paper 639, presented at Corrosion/2000, NACE International, 2000 43. R.M. Kain et al., “Crevice Corrosion of Nickel-Chromium-Molybdenum Alloys in Natural and Chlorinated Seawater,” Paper 112, presented at Corrosion/89, National Association of Corrosion Engineers, 1989 44. R.M. Kain and P.A. Klein, “Crevice Corrosion Propagation Studies for Alloy N06625: Remote Crevice Assembly Testing in Flowing Natural and Chlorinated Seawater,” Paper 158, presented at Corrosion/90, National Association of Corrosion Engineering, 1990 45. R.M. Kain and A. Zeuthen, Crevice Corrosion Testing of Austenitic, Superaustenitic, Superferritic and Superduplex Stainless Type Alloys in Seawater, Corrosion Testing in Natural Waters, STP 1300, American Society for Testing and Materials, 1996 46. M.E. Inman et al., “Detection of Crevice Corrosion in Natural Seawater Using Polarization Resistance Measurements,” Paper 297, presented at Corrosion/97, NACE International, 1997 47. D.C. Agarwal, “Solving Critical Problems in Marine Environments by an Advanced Ni-Cr-Mo Alloy 59 UNS N06059,” Paper 635, presented at Corrosion/2000, NACE International, 2000 48. R.M. Kain, “Testing for Crevice Corrosion Susceptibility,” presented at research symposium at Corrosion/96, NACE International, 1996 49. LaQue Center for Corrosion Technology, Inc., Unpublished test report, 1993 50. R.M. Kain and M.B. Ives, “Surface Treatments Benefiting the Crevice Corrosion Resistance of Nickel Alloy N06625 in Natural and Chlorinated Seawater,” Paper 330, presented at Corrosion/99, NACE International, 1999 51. R.M. Kain et al., “Localized and General Corrosion Resistance of Candidate Metallic Materials for ROMembrane Cartridges,” Paper 268, presented at Corrosion/95, NACE International, 1995 52. R.M. Kain, A. Zeuthen, and J. Maurer, “Localized Corrosion Resistance of Stainless Type Materials in Aerated, Deaerated, and Stagnant Sulfide Bearing Seawater,” Paper 423, presented at Corrosion/97, NACE International, 1997 53. P.A. Klein et al., “A Localized Corrosion Assessment of 6% Molybdenum Stainless Steel Condenser Tubing at the Calvert Cliffs Nuclear Power Plant,” Paper 490, presented at Corrosion/94, National Association of Corrosion Engineers, 1994 54. R.M. Kain, “Crevice Corrosion Behavior of Coated Stainless Steel in Natural Seawater,” Paper 827, presented at Corrosion/2000, NACE International, 2000 55. R.M. Kain, “Seawater Crevice Corrosion Resistance of Stainless Steels Coated with Silane and Antifouling Paint Systems,” Paper 187, presented at Corrosion/2002, NACE International, 2002 56. LaQue Center for Corrosion Technology, Inc., Unpublished test report, 1998

57. E.L. Hibner and L.E. Shoemaker, “High Strength Corrosion Resistance Alloy 686 for Seawater Fastener Service,” Paper 195, presented at Corrosion/2002, NACE International, 2002 58. “Standard Practice for Conducting Cyclic Potentiodynamic Polarization Measurements for Localized Corrosion,” G 61, Annual Book of ASTM Standards, Vol 03.02, American Society for Testing and Materials 59. R.M. Kain, “Localized Corrosion Behavior in Natural Seawater: A Comparison of Electrochemical and Crevice Testing of Stainless Steel,” Paper 70, presented at Corrosion/80, National Association of Corrosion Engineers, 1980 60. B.E. Wilde, A Critical Appraisal of Some Popular Laboratory Electrochemical Tests for Predicting the Localized Corrosion Resistance of Stainless Alloys in Sea Water, Corrosion, Vol 28 (No. 8), 1972, p 283 61. S. Bernhardsson, Paper 85, presented at Corrosion/80, National Association of Corrosion Engineers, 1980 62. N.S. Nagaswami and M.A. Streicher, “Accelerated Laboratory Tests for Crevice Corrosion of Stainless Alloys,” Paper 7, presented at Corrosion/83, National Association of Corrosion Engineers, 1983 63. J.M. Kroughman and F.P. Ijsseling, Crevice Corrosion of Stainless Steels and Nickel Alloys in Seawater, Proc. Fifth International Congress on Marine Corrosion and Fouling, G. Londres, Ed., (Barcelona, Spain), 1980, p 214 64. A. Mollica et al., Cathodic Performance of Stainless Steels in Natural Seawater as a Function of Microorganism Settlement and Temperature, Corrosion, Vol 48 (No. 1), 1998, p 48–56 65. T.S. Lee et al., Use of Electrochemical Techniques for the Study of Crevice Corrosion in Natural Seawater, Electrochemical Techniques for Corrosion Engineering, National Association of Corrosion Engineering, 1986 66. “Standard Test Method for Electrochemical Critical Pitting Temperature Testing of Stainless Steels,” G 150, Wear and Erosion, Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards, American Society for Testing and Materials 67. T.S. Lee, A Method of Quantifying the Initiation and Propagation Stages of Crevice Corrosion, Electrochemical Corrosion Testing, STP 727, American Society for Testing and Materials, 1981, p 43– 68 68. R.M. Kain, Electrochemical Measurement of the Crevice Corrosion Propagation Resistance of Stainless Steels: Effect of Environmental Variables, Mater. Perform., Vol 23 (No. 2), 1984, p 24 69. R.M. Kain and T.S. Lee, “The Effect of Crevice Solution pH on Corrosion Behavior of Stainless Steels,” Paper 27, presented at Corrosion/84, National Association of Corrosion Engineers, 1984 70. H.S. Issacs, “Application of the Vibration Probe to Localized Current Measurements,” Paper 55, presented at Corrosion/85, National Association of Corrosion Engineers, 1985 71. J. Postlewaite, Can. Metall. Q., Vol 22 (No. 1), 1983, p 133 72. F.P. Ijsseling, Electrochemical Methods in Crevice Corrosion Testing, Br. Corros. J., Vol 15 (No. 2), 1980, p 51

73. J.W. Oldfield and R.M. Kain, Prediction of Crevice Corrosion Resistance of Stainless Steels in Aqueous Environments—A Corrosion Engineering Guide, Proc. 12th International Corrosion Congress, National Association of Corrosion Engineers, 1993, p 1876 74. P.O. Gartland, Modelling Crevice Corrosion of Fe-Cr-Ni-Mo Alloys in Chloride Solutions, Proc. 12th International Corrosion Congress, Vol 3B, National Association of Corrosion Engineers, 1993, p 1901– 1914 75. B. Shaw, P. Moran, and P.O. Gartland, Crevice Corrosion of a Nickel-Based Super Alloy in Natural and Chlorinated Seawater, Proc. 12th International Corrosion Congress, Vol 3B, National Association of Corrosion Engineers, 1993, p 1915–1928



F.L. LaQue, Marine Corrosion—Causes and Prevention, Crevice Corrosion, J. Wiley & Sons, 1975, p 164 A.J. Sedriks, Corrosion of Stainless Steel, Crevice Corrosion, 2nd ed., Wiley Interscience, 1996

H.P. Hack, Evaluation Galvanic Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 562–567

Evaluating Galvanic Corrosion Harvey P. Hack, Northrop Grumman Corporation

Introduction GALVANIC CORROSION, although listed as one of the forms of corrosion, should instead be considered a type of corrosion mechanism, because galvanic effects can accelerate any of the other forms of corrosion. Any of the tests used for the more conventional forms of corrosion, therefore, such as uniform attack, pitting, or stress corrosion, can be used with modifications to determine galvanic-corrosion effects. The modifications can be as simple as connecting a second metal to the system or as complex as necessary to evaluate the appropriate parameters. A change in the method of data interpretation is often all that is needed to convert conventional test methods into galvanic-corrosion tests. This article discusses component, model, electrochemical, and specimen tests.

H.P. Hack, Evaluation Galvanic Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 562–567 Evaluating Galvanic Corrosion Harvey P. Hack, Northrop Grumman Corporation

Component Testing Component testing is an especially useful technique for galvanic corrosion prediction. The materials in a system are often selected primarily for reasons other than galvanic compatibility. In complex components, such as valves or pumps, many different materials can be used in a geometric configuration that is extremely difficult to model. In more complicated cases, even the most basic prediction, such as which materials will suffer increased corrosion due to galvanic effects, may not be possible from simple laboratory tests. Therefore, component testing becomes the best method for predicting galvanic corrosion behavior in complex systems. Conducting component tests for galvanic corrosion is similar to conducting component tests for any other type of corrosion. The same care must be taken to ensure that the materials, the operation of the component, and the environment are similar to those in service. However, one important difference with regard to galvanic corrosion is the relationship between the component being tested and the other elements of the system. It would, for example, be a waste of effort to expose a complicated piece of machinery in order to look for galvanic corrosion when the whole device is cathodically protected as a result of being attached to a protected structure. Alternatively, incorrect results would be obtained for the exposure of an isolated bronze mixedmaterial valve when the ultimate use is in a piping system made of a more noble metal that could accelerate the corrosion of the entire valve galvanically. When outside interactions of this type are possible, the interacting materials must be made part of the corrosion system by exposing the appropriate surface area of those materials electrically connected to, and in the same electrolyte as, the component being tested. The principal advantages of component testing are the ease of interpretation of results and the lack of scaling or modeling uncertainties. The disadvantages include the high cost and the need for extremely sensitive measures of corrosion damage to obtain results within reasonable time periods.


Modeling Even when the galvanic behavior of panels of the materials of interest is known, the geometrical arrangement of these materials may make galvanic-corrosion prediction difficult because of the effects of solution (electrolyte) resistance on the corrosion rates. An example of this is a heat-exchanger tube in a tubesheet. Assuming the tube to be anodic to the tubesheet, areas of the tube near the tubesheet will have low solution resistance to the cathode and will corrode rapidly, but areas away from the tubesheet will have a large solution resistance to the cathodic metal and will therefore corrode more slowly. In the reverse case, in which the tubesheet is anodic to the tube, the areas of the cathodic tube near the tubesheet will drive the galvanic corrosion of the tubesheet much more than distant areas will. Computer Modeling. Geometrical effects can be modeled in computers by using such techniques as finite elements, boundary elements, and finite differences. The best computer models solve a version of the Laplace

equation for the electrolyte surrounding the corroding materials and use the polarization behavior of the material in question as boundary conditions at the metal/ electrolyte interface. The analysis is similar to the heat flow analysis, with potential analogous to temperature, current analogous to heat flux, and the polarization boundary condition analogous to a special nonlinear type of temperature- dependent convective flux. The only data this type of model requires are the geometry, electrolyte conductivity, and polarization characteristics of the materials involved. The program generates potentials and current densities as a function of location, either of which can be related back to corrosion rate. The nonlinear boundary conditions make this type of computer modeling difficult to perform unless sufficient computing power is available. Computer modeling provides an excellent predictive tool for geometrical effects; however, it is still seen as less satisfying than physical scale- model exposures. Physical scale modeling must model the solution resistance effects and the relative effects of polarization resistance and solution resistance to obtain accurate geometrical predictive capability. When solution resistance is important, the best type of scale modeling is the scaled conductivity exposure. In this type of exposure, the model is reduced in size by some factor from the original. To maintain proper potential and current distribution scaling, the electrolyte conductivity must also be reduced by the same factor. Any resistive coatings such as paints must also have their conductivity scaled similarly. In the case of paints, this can be accomplished by applying a thinner layer, reduced by the same scaling factor used for size, than the thickness used in practice. For example, a one-tenth-scale model of a heat exchanger designed to operate in seawater with a conductivity of 4 S/cm should be placed in seawater diluted to a conductivity of 0.4 S/cm. In this case, the observed potential and current distributions will be the same between the model and the full-scale heat exchanger. For physical scale modeling, measurements that can be taken include potential distribution by the use of a movable reference electrode, corrosion depth as a function of location, and, if the model design permits, current to different parts of the structure and mass loss of certain model components. Although less expensive than full-scale component testing, physical scale modeling has many of the disadvantages of component testing. In addition, a great inaccuracy in conductivity scaling stems from the fact that the polarization resistance of the materials in the system of interest is often a function of solution conductivity. Thus, changing solution conductivity may influence polarization resistance sufficiently to make the results of the model inaccurate. There is no experimental way to avoid this shortcoming.


Laboratory Testing Laboratory tests fall into two categories: electrochemical tests in which the data are analyzed and reported in a way that assists galvanic-corrosion predictions and specimen exposures that may or may not be electrochemically monitored.

Electrochemical Tests The use of electrochemical techniques to predict galvanic corrosion is summarized in Ref 1. The details that relate to testing techniques are discussed in the following paragraphs. Galvanic Series. When the only information needed is which of the materials in the system are possible candidates for galvanically accelerated corrosion and which will be unaffected or protected, the information from a galvanic series in the appropriate medium is useful. Such a series is a list of freely corroding potentials of the materials of interest in the environment of interest arranged in order of potential (Fig. 1). The galvanic series is easy to use and is often all that is required to answer a simple galvanic-corrosion question. The

material with the most negative, or anodic, corrosion potential has a tendency to suffer accelerated corrosion when electrically connected to a material with a more positive, or noble, potential. The disadvantages include: • • • •

No information is available on the rate of corrosion. Active-passive metals may display two widely differing potentials. Small changes in electrolyte can change the potentials significantly. Potentials may be time dependent.

Fig. 1 Galvanic series for seawater. Dark boxes indicate active behavior of active-passive alloys.

Creating a galvanic series is a matter of measuring the corrosion potential of various materials of interest in the electrolyte of interest against a reference electrode half-cell, such as saturated calomel. This procedure is described in Ref 2. The details of such factors as meter resistance, reference cell selection, and measurement duration are also addressed in Ref 2. There is little difference from a normal reading of corrosion potential except for the measurement duration and the creation of a list ordered by potential. To prepare a galvanic series that will be valid for the materials and environment of interest in service, all of the factors that affect the potential of those materials in that environment must be accounted for. These factors include material composition, heat treatment, surface preparation (mill scale, coatings surface finish, etc.), environmental composition (trace contaminants, dissolved gases, etc.), temperature, and flow rate. One important effect is exposure time, particularly on materials that form corrosion product layers. All of the precautions and warnings regarding the generation and use of a galvanic series are given in Ref 2. Polarization Curves. More useful information on the rate of galvanic corrosion can be obtained by investigating the polarization behavior of the materials involved. This can be done by generating stepped potential or potentiodynamic polarization curves or by obtaining potentiostatic information on polarization behavior. The objective is to obtain a good indication of the amount of current required to hold each material at a given potential. Because all materials in the galvanic system must be at the same potential in systems with low solution resistivity, such as seawater, and because the sum of all currents flowing between the materials must equal 0 by Kirchoff's Law, the coupled potential of all materials and the galvanic currents flowing can be uniquely determined for the system. Faraday's Law can then be used to relate the corrosion rate to the galvanic current if the resulting potential of the anodic materials is well away from their corrosion potential, or the corrosion rate can be found as a function of potential by independent measurement. Potentiodynamic polarization curves are generated by connecting the specimen of interest to a scanning potentiostat. This device applies whatever current is necessary between the specimen and a counterelectrode to maintain that specimen at a given potential versus a reference electrode half-cell placed near the specimen. The current required is plotted as a function of potential over a range that begins at the corrosion potential and proceeds in the direction (anodic or cathodic) required by that material. Such curves would be generated for each material of interest in the system. Additional information on the method for generating these curves is available in the article “Electrochemical Methods of Corrosion Testing” in this Volume and in Ref 3. The scan rate for potential must be chosen such that sufficient time is allowed for completion of electrical charging at the interface. Potentiodynamic polarization is particularly effective for materials with time-independent polarization behavior. It is fast, relatively easy, and gives a reasonable, quantitative prediction of corrosion rates in many systems. However, potentiostatic techniques are preferred for time-dependent polarization. To establish polarization characteristics for time-dependent polarization, a series of specimens is used, each held to one of a series of constant potentials with a potentiostat while the current required is monitored as a function of time. After the current has stabilized or after a preselected time period has elapsed, the current at each potential is recorded. Testing of each specimen results in the generation of one potential/current data pair, which gives a point on the polarization curve for that material. The data are then interpolated to trace out the full curve. This technique is very accurate for time- dependent polarization but is expensive and time consuming. The individual specimens can be weighed before and after testing to determine corrosion rate as a function of potential, thus enabling the errors from using Faraday's Law to be easily corrected. The process of predicting galvanic corrosion from polarization behavior can be illustrated by the example of a steel-copper system. Steel has the more negative corrosion potential and will therefore suffer increased corrosion on coupling to copper, but the amount of this corrosion must be predicted from polarization curves. If the polarization of each material is plotted as the absolute value of the log of current density versus potential and if the current density axis of each of these curves is multiplied by the wetted surface area of that material in the service application, then the result will be a plot of the total anodic current for steel and the total cathodic current for copper in this application as a function of potential (Fig. 2).

Fig. 2 Prediction of coupled potential and galvanic current from polarization diagrams. i, current; io, exchange current; Ecorr, corrosion potential Furthermore, when the two metals are electrically connected, the anodic current to the steel must be supplied by the copper; that is, the algebraic sum of the anodic and cathodic currents must equal 0. If the polarization curves for the two materials, normalized for surface area as described previously, are plotted together, this current condition is satisfied where the two curves intersect. This point of intersection allows for the prediction of the coupled potential of the materials and the galvanic current flowing between them from the intersection point. This procedure works if there is no significant electrolyte resistance between the two metals; otherwise, this resistance must be taken into account as described in Ref 4.

Specimen Exposures Specimens for galvanic-corrosion testing include panels, wires, pieces of actual components, and other configurations of the materials of interest that are exposed in a process stream, a simulated service environment, or the actual environment. Specimens of the materials of interest are usually exposed in the same ratios of wetted or exposed areas as in the service application. The different materials are either placed in physical contact to provide electrical connection or are wired together such that the current between the materials can be monitored, usually as a function of time. Seldom can the effects of electrolyte resistance be included in this type of test, and the resistance is usually kept extremely low by appropriate relative placement of the materials. Immersion. There are virtually no standardized tests for galvanic corrosion under immersion conditions, partly because the type of information needed, the extent of modeling of the service situation, and the type of system studied vary widely. This makes development of a standard test difficult. However, some general guidelines for galvanic-corrosion specimen testing in liquid electrolytes are given in Ref 5. Immersion testing always involves an electrical connection between at least two dissimilar metals. This is usually accomplished with a wire, as in Fig. 3, although threaded mounting rods such as the assembly shown in Fig. 4 have also been used successfully for electrical connection. Soldered or brazed connections have the best electrical integrity.

Fig. 3 Typical galvanic-corrosion immersion test setup using wire connections

Fig. 4 Typical galvanic-corrosion test specimen using a threaded rod for mounting and electrical connection The electrolyte must be excluded from the contact area by applying a sealant, such as silicone or epoxy; by keeping the joint area out of the electrolyte by partial immersion of the specimen, in which case a waterline area is created; or by use of a tube and gasket or O-ring seal, in the case of a threaded mounting rod. Mounting the specimen in a specially formulated epoxy has been found to be effective in minimizing crevice corrosion while maintaining a dry electrical connection. However, selection of the best epoxy formulation is difficult. Care must be taken that the sealant or gasket material is stable in the electrolyte being studied. Almost any sealing procedure will create a potential area for crevice corrosion; thus, it is very difficult to study galvanic behavior independent of crevice-corrosion behavior. Control specimens may be run with similar crevices and no electrical connection, but because the reproducibility of crevice-corrosion behavior is not good, data scatter will be large. Under some circumstances, the galvanic effect of importance may be the acceleration of crevice-corrosion attack. The relative wetted surface areas of the materials being tested will have an important effect on the magnitude of the galvanic attack. The larger the cathode-to-anode area ratio is, the larger the degree of attack will be; therefore, it would at first seem reasonable to accelerate the test by using a large ratio. This should not be done, because accelerating the attack may also change the mechanism of the attack, which would lead to erroneous conclusions. It is far more appropriate to use more accurate measurement techniques to determine the extent of

the attack over a short period than to accelerate the test to obtain measurable attack quickly. If soldered or brazed connections are used for electrical connection, subsequent evaluation by weight loss is difficult; therefore, if weight loss is to be used to measure attack, threaded and sealed connections are preferred. Measurement of the electrical current flowing between the metals can give a very sensitive indication of the extent of the galvanic attack and will allow the attack to be monitored over time. Coupled potential is another parameter that is useful to follow during the course of the exposure. The effect of exposure time on the rate of attack should be properly considered. Initially high rates of galvanic attack may decay to acceptable levels in a short period of time, or initially low attack rates may increase to unacceptable levels over time. Current can be measured by inserting a resistor of 1 to 10 Ω in the current circuit and measuring the potential drop across this resistor with a voltmeter having an internal resistance of at least 1000 Ω. The resistor should be sized such that the voltage across it does not exceed 5 mV; thus, the resistor will not significantly impede the current flow. Alternatively, a zero-resistance ammeter (ZRA) can be used instead of the resistor. This device is an operational amplifier connected to maintain 0 V across its input terminals (Fig. 5). A current-measuring resistor placed in the feedback circuit may be as large as the amplifier will allow, thus enabling currents in the nanoampere range to be easily measured. One simple way of creating a zero-resistance ammeter is by using a potentiostat with the counterelectrode and reference electrode leads shorted together and set to a working electrode potential of 0 V (Fig. 6).

Fig. 5 Basic circuit for a zero-resistance ammeter

Fig. 6 Conversion of a potentiostat into a zero-resistance ammeter. WE, working electrode; CE, counter electrode; RE, reference electrode For most electrochemical reactions it is possible to convert reaction current to corrosion rate using the expression: CR = 0.1288igEW/d where CR is corrosion rate in mils per year, ig is current density in microamperes per square centimeter, EW is equivalent weight of the corroding material, and d is density in grams per cubic centimeter. This conversion will frequently be inaccurate for calculating galvanic corrosion rate from galvanic current because corrosion may not be uniform, some current may go toward reactions at the anode other than metal loss, such as valence change of ions in solution, and additional corrosion at the anode is generated by cathodic reactions on the anode not measured by the galvanic current. Therefore, the preceding calculation will frequently underestimate the total corrosion of an anode in a galvanic corrosion cell. Estimation of galvanic corrosion rate from galvanic

current is worse for materials with high self-corrosion, such as metals in acids, and best when the anode and cathode open-circuit potentials are far apart so that after coupling, little cathodic activity occurs on the anode. The importance of electrolyte flow in galvanic corrosion should not be overlooked in establishing the test procedure. A test apparatus should be used that reproduces the flow under service conditions. If this is not possible and flow must be scaled, the exact scaling method will depend on the assumed corrosion processes. Cathodic reactions, such as oxygen reduction, that are controlled by diffusion through a fluid boundary layer are likely to be properly scaled by reproducing the hydrodynamic boundary layer of the service application in the test. This should reproduce the diffusion boundary layer that controls the reaction. Alternatively, the rates of reactions controlled by films such as anodic brightening of copper alloys, other passivation-type reactions, or control by calcareous-deposit formation in seawater, may depend more on the shear stress at the surface required to strip off the film. In this case, surface shear stress may be a better hydrodynamic parameter to reproduce in the test. Many different types of flow apparatus have been used, such as concentric tubes, in-line tubes, rotating cylinders, rotating ring-disks, rectangular flow channels with specimens mounted in the walls, and circulating foils. Each of these has its own hydrodynamic peculiarities, but one common area of concern is the leading edge of the specimen. It is difficult, even for specimens recessed in the walls of a flow channel, to avoid a step or gap that can create unexpected hydrodynamic conditions at the specimen surfaces downstream. Also, mounting to allow electrical connection must be considered, and crevice effects are essentially impossible to eliminate. Atmospheric Tests. General testing guidelines become more complex when considering atmospheric or cabinet exposures. Testing in these environments differs markedly from immersion tests in a number of ways, most of which involve the insufficiency of electrolyte. Many of the variables that influence the behavior of specimens in the atmosphere are discussed in Ref 6. The thinness of the electrolyte film and the normally low conductivity of the electrolyte combine to limit galvanic effects for bare metals to within approximately 5 mm (0.2 in.) of the dissimilar-metal interface (Ref 7). This distance may be somewhat greater if nonconductive coatings are present. Area ratio effects thus become relatively unimportant. Sealing the electrical connections becomes relatively less important than in immersion testing if the connections are more than 5 mm (0.2 in.) from the area to be evaluated and if corrosion products will not interfere with the continuity of the connection. Periodic checks of electrical continuity in atmospheric galvanic-corrosion tests are recommended. Geometrical effects also become unimportant, except as they relate to the entrapment of moisture. However, specimen orientation effects must be considered. The behavior of the galvanic couples will depend on whether they are exposed on the top or the bottom of the panel, whether they are sheltered or not, or other considerations, such as the effect of specimen mass on condensation. It is surprising that several standard tests have emerged for atmospheric galvanic corrosion, since there are no standardized tests for galvanic corrosion immersed in electrolytes even though more testing has been done in this area. One of these tests is an International Organization for Standardization (ISO) standard (Ref 8). This test uses a 100 × 400 mm (4 × 16 in.) panel of the anodic material to which a 50 × 100 mm (2 × 4 in.) strip of the cathodic material is bolted (Fig. 7). After exposure, the anodic material can be evaluated for material degradation by weight loss and other corrosion measurements as well as by degradation of such mechanical properties as ultimate tensile strength.

Fig. 7 Specimen configuration for the ISO test for atmospheric galvanic corrosion. 1, anodic plate, 1 piece; 2, cathodic plate, 2 pieces; 3, microsection, 2 pieces; 4, tensile test specimen; 5, bolt, 8 × 40 mm, 2 pieces; 6, washers, 1 mm thick, 16 mm diameter, 4 pieces; 7, insulating washers, 1 to 3 mm thick, 18 to 20 mm diameter, 4 pieces; 8, insulating sleeve, 2 pieces; 9, nut, 2 pieces. This test is relatively easy to perform but requires the availability of plates of the materials of interest and a prior knowledge of which material is anodic. Like any atmospheric galvanic- corrosion test, crevice effects cannot be adequately separated from galvanic effects in some cases; therefore, a coating is sometimes applied between the anode and cathode plates. The disadvantage of this test is the time required to obtain results; for systems with moderate corrosion rates, exposures of 1 to 5 years are not unusual. Another commonly used atmospheric galvanic-corrosion test is the wire-on-bolt test, sometimes referred to as the CLIMAT test (Ref 9, 10, 11). In this test, a wire of the anodic material is wrapped around a threaded rod of the cathodic material (Fig. 8). Because corrosion can be rapid in this test, exposure duration should usually be limited. This makes the test ideal for measuring atmospheric corrosivity as well as material corrosion properties. Not all materials of interest are available in the required wire and threaded rod forms, and analysis is usually restricted to weight-loss measurement and observation of pitting. When the required materials are available, this test is less expensive and easier to conduct than the ISO plate test.

Fig. 8 Specimen configuration for the wire-on-bolt test for atmospheric galvanic corrosion A third atmospheric galvanic-corrosion test has been used extensively by ASTM International but has not been standardized. This test (Ref 12) involves the use of 25 mm (1 in.) diameter washers of the materials of interest bolted together as shown in Fig. 9. The bolt that holds the washers together can also be used to secure the assembly in position. After exposure, the washers can be disassembled for weight loss determination. The materials needed for this test are not as large as those for the ISO plate test, but testing can take as long and cannot provide mechanical properties data.

Fig. 9 Specimen configuration for the washer test for atmospheric galvanic corrosion

References cited in this section 1. R. Baboian, Electrochemical Techniques for Predicting Galvanic Corrosion, Galvanic and Pitting Corrosion—Field and Laboratory Studies, STP 576, ASTM International, 1976, p 5–19 2. “Standard Guide for Development and Use of a Galvanic Series for Predicting Galvanic Corrosion Performance,” G 82, Annual Book of ASTM Standards, ASTM International 3. “Standard Reference Test Method for Making Potentiostatic and Potentiodynamic Anodic Polarization Measurements,” G 5, Annual Book of ASTM Standards, ASTM International 4. H.P. Hack, P.J. Moran, and J.R. Scully, Influence of Electrolyte Resistance on Electrochemical Measurements and Procedures to Minimize or Compensate for Resistance Errors, The Measurement and Correction of Electrolyte Resistance in Electrochemical Tests, STP 1056, L.L. Scribner and S.R. Taylor, Ed., ASTM International, Jan 1990, p 5–26 5. “Standard Guide for Conducting and Evaluating Galvanic Corrosion Tests in Electrolytes,” G 71, Annual Book of ASTM Standards, ASTM International 6. “Standard Practice for Conducting Atmospheric Corrosion Tests of Metals,” G 50, Annual Book of ASTM Standards, ASTM International 7. V. Kucera and E. Mattson, Atmospheric Corrosion of Bimetallic Structures in Atmospheric Corrosion, W.H. Aylor, Ed., John Wiley & Sons, 1982, p 567 8. “Corrosion of Metals and Alloys—Determination of Bi-Metallic Corrosion in Outdoor Exposure Corrosion Tests,” ISO 7441, International Organization for Standards 9. K.G. Compton, A. Mendizza, and W.W. Bradley, Atmospheric Galvanic Couple Corrosion, Corrosion, Vol II, 1955, p 383 10. H.P. Godard, Galvanic Corrosion Behavior of Aluminum in the Atmosphere, Mater. Prot., Vol 2 (No. 6), 1963, p 38 11. D.P. Doyle and T.E. Wright, Rapid Methods for Determining Atmospheric Corrosivity and Corrosion Resistance, Atmospheric Corrosion, W.H. Aylor, Ed., John Wiley & Sons, 1982, p 227 12. R. Baboian, Final Report on the ASTM Study: Atmospheric Galvanic Corrosion of Magnesium Coupled to Other Metals, Atmospheric Factors Affecting the Corrosion of Engineering Metals, STP 646, S.K. Coburn, Ed., ASTM International, 1978, p 17–29


References 1. R. Baboian, Electrochemical Techniques for Predicting Galvanic Corrosion, Galvanic and Pitting Corrosion—Field and Laboratory Studies, STP 576, ASTM International, 1976, p 5–19 2. “Standard Guide for Development and Use of a Galvanic Series for Predicting Galvanic Corrosion Performance,” G 82, Annual Book of ASTM Standards, ASTM International 3. “Standard Reference Test Method for Making Potentiostatic and Potentiodynamic Anodic Polarization Measurements,” G 5, Annual Book of ASTM Standards, ASTM International 4. H.P. Hack, P.J. Moran, and J.R. Scully, Influence of Electrolyte Resistance on Electrochemical Measurements and Procedures to Minimize or Compensate for Resistance Errors, The Measurement and Correction of Electrolyte Resistance in Electrochemical Tests, STP 1056, L.L. Scribner and S.R. Taylor, Ed., ASTM International, Jan 1990, p 5–26 5. “Standard Guide for Conducting and Evaluating Galvanic Corrosion Tests in Electrolytes,” G 71, Annual Book of ASTM Standards, ASTM International 6. “Standard Practice for Conducting Atmospheric Corrosion Tests of Metals,” G 50, Annual Book of ASTM Standards, ASTM International 7. V. Kucera and E. Mattson, Atmospheric Corrosion of Bimetallic Structures in Atmospheric Corrosion, W.H. Aylor, Ed., John Wiley & Sons, 1982, p 567 8. “Corrosion of Metals and Alloys—Determination of Bi-Metallic Corrosion in Outdoor Exposure Corrosion Tests,” ISO 7441, International Organization for Standards 9. K.G. Compton, A. Mendizza, and W.W. Bradley, Atmospheric Galvanic Couple Corrosion, Corrosion, Vol II, 1955, p 383 10. H.P. Godard, Galvanic Corrosion Behavior of Aluminum in the Atmosphere, Mater. Prot., Vol 2 (No. 6), 1963, p 38 11. D.P. Doyle and T.E. Wright, Rapid Methods for Determining Atmospheric Corrosivity and Corrosion Resistance, Atmospheric Corrosion, W.H. Aylor, Ed., John Wiley & Sons, 1982, p 227 12. R. Baboian, Final Report on the ASTM Study: Atmospheric Galvanic Corrosion of Magnesium Coupled to Other Metals, Atmospheric Factors Affecting the Corrosion of Engineering Metals, STP 646, S.K. Coburn, Ed., ASTM International, 1978, p 17–29



R. Francis, Galvanic Corrosion: A Practical Guide for Engineers, NACE International, 2001 Galvanic Corrosion, STP 978, H.P. Hack, Ed., ASTM International, 1988 H.P. Hack, Galvanic Corrosion Test Methods, NACE International, 1993

B. Phull, Evaluating Intergranular Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 568–571

Evaluating Intergranular Corrosion Revised by Bopinder Phull, Consultant

Introduction INTERGRANULAR CORROSION (IGC) is preferential attack of either grain boundaries or areas immediately adjacent to grain boundaries in a material exposed to a corrosive environment but with little corrosion of the grains themselves. Intergranular corrosion is also known as intergranular attack (IGA). End-grain attack, grain dropping, and “sugaring” are additional terms that are sometimes used to describe IGC. In certain materials, corrosion progressing laterally along planes parallel to rolled surfaces is known as exfoliation, and it generally occurs along grain boundaries—hence, intergranular corrosion. A layered appearance is a common manifestation of exfoliation (also referred to as layer corrosion), resulting from voluminous corrosion products prying open the material; for example, in aluminum alloys. Most alloys are susceptible to IGA when exposed to specific environments. This is because grain boundaries are sites for precipitation and segregation, which make them chemically and physically different from the grains themselves. Intergranular attack is defined as the selective dissolution of grain boundaries or closely adjacent regions without appreciable attack of the grains themselves. This is caused by potential differences between the grain-boundary region and any precipitates, intermetallic phases, or impurities that form at the grain boundaries. The actual mechanism and degree of attack differs for each alloy system. Precipitates that form as a result of the exposure of metals at elevated temperatures (for example, during production, fabrication, heat treatment, and welding) often nucleate and grow preferentially at grain boundaries. If these precipitates are rich in alloying elements that are essential for corrosion resistance, the regions adjacent to the grain boundary are consequently depleted of these elements. The metal is thus sensitized and is susceptible to IGA in one or more specific corrosive environments. For example, in austenitic stainless steels such as Type 304, intergranular attack is often associated specifically with precipitation of chromium-rich carbides (Cr23C6) at grain boundaries in the heat- affected zone. Precipitation of such carbides is often referred to as sensitization. When chromium-rich precipitates form, the surrounding areas are depleted in chromium. As a result, the depleted areas are more susceptible to corrosion in specific environments than regions away from the grain boundaries. Another example of grain boundary segregation is sigma-phase formation that results in a

Cr- and Mo-rich constituent at grain boundaries in Cr- and Mo-containing alloys. Sigma-phase is usually more difficult to resolve visually in the microstructure than Cr-carbides. Impurities that segregate at grain boundaries may promote galvanic action in a corrosive environment by serving either as anodic or cathodic sites. For example, in 2000-series (2xxx) aluminum alloys, the copperdepleted (anodic) band on either side of the grain boundary is dissolved while the grain boundary is cathodic due to the CuAl2 precipitates. Conversely, in the 5000-series (5xxx) aluminum alloys, intermetallic precipitates such as Mg2Al3 (anodic) are attacked when they form a continuous phase in the grain boundary. During exposures to chloride solutions, the galvanic couples formed between these precipitates and the alloy matrix can lead to severe intergranular attack. Actual susceptibility to intergranular attack and degree of corrosion depends on the corrosive environment and on the extent of intergranular precipitation, which is a function of alloy composition, fabrication, and heat treatment parameters.

B. Phull, Evaluating Intergranular Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 568–571 Evaluating Intergranular Corrosion Revised by Bopinder Phull, Consultant

The Purpose of Testing There is a perception in much of the industry that testing for susceptibility to IGA is equivalent to evaluating the resistance of the alloy to general and localized corrosion. Although the tests used for evaluating susceptibility to IGA are severe, they are not intended to duplicate conditions for the wide range of chemical exposures present in an industrial plant, even though some of these tests do simulate service conditions. Testing for susceptibility to IGA, however, is useful for determining whether a vendor has supplied the correct material and in the proper metallurgical condition. There are some problems associated with quality assurance programs for purchased materials. Such programs are sometimes based on faith in what is supplied by a vendor or production mill and what is certified in the documentation sent with the material. However, such confidence may be misplaced. For example, there have been a number of accounts in which alloys have been substituted, resulting in premature failure. In one case, this occurred when Hastelloy B (UNS N10001) valves were substituted for the Hastelloy C-276 (UNS N10276) valves that were ordered to handle a hypochlorite solution. Not surprisingly, the Hastelloy B valves failed in about 3 months because this alloy is usually specified for reducing environments (e.g., HCl), whereas alloy C-276 is typically more suited to oxidizing environments (e.g., hypochlorite). In addition, there are many examples in which the material supplied does not conform to its certified analysis. The problem of getting reliable certified analyses increases when documentation goes from a mill to an alloy supplier. In one case, for example, Type 304L stainless steel (UNS S30403) valves were ordered, but the vendor, having few orders for this alloy, substituted Type 316L stainless steel (UNS S31603) valves and sent certifications that purposely omitted the molybdenum analysis. Normally, this would have been a good substitution for improved corrosion resistance at a bargain price, but these valves were destined for hot, concentrated HNO3 service and failed prematurely. Therefore, it is essential that, for critical service, the correct alloys must be specified and in optimum metallurgical condition to resist IGA and other forms of corrosion associated with precipitates at the grain boundaries.


Tests for Stainless Steels and Nickel-Base Alloys The austenitic and ferritic stainless steels, as well as most nickel-base alloys, are generally supplied in a heat treated condition such that they are free of carbide precipitates that are detrimental to corrosion resistance. However, these alloys are susceptible to sensitization from welding, improper heat treatment, and service in the sensitizing temperature range. The theory and application of acceptance tests for detecting the susceptibility of stainless steels and nickel-base alloys to intergranular attack are extensively reviewed in Ref 1, 2, 3. Corrosion tests for evaluating the susceptibility of an alloy to IGA are typically classified as either simulated-service or accelerated tests. The original laboratory tests for detecting IGA were simulated-service exposures. These were first used in 1926 when IGA was detected in an austenitic stainless steel in a copper sulfate-sulfuric acid (CuSO4-H2SO4) pickling tank (Ref 4). Another simulated-service test for alloys intended for service in nitric acid (HNO3) plants is described in Ref 5. In this case, for accelerated results, iron-chromium alloys were tested in a boiling 65% HNO3 solution. Over the years, specific accelerated tests have been developed and standardized for evaluating the susceptibility of various alloys to IGA. For example, ASTM A 262 contains six practices for detecting susceptibility to IGC in austenitic stainless steels (Ref 6). Practice A is an electrolytic oxalic-acid-etch screening test that can be performed in minutes on prepared samples followed by microscopic examination of the etched microstructure for Cr23C6 sensitization. Practice B is a 120 hour test in a boiling solution of ferric sulfate [Fe2(SO4)3] + sulfuric acid (H2SO4); the weight-loss corrosion rate indicates degree of IGA due to Cr23C6 precipitation. A 240 hour test in boiling 65% HNO3 constitutes Practice C; the degree of IGC due to Cr23C6 and sigma-phase formation is indicated by weight-loss corrosion rates. Practice D, which was an immersion test in 10% HNO3 + 3% HF solution at 70 °C has been removed from ASTM A 262 (Ref 6). Practice E is a 24 hour test in boiling 6% CuSO4 + 10% H2SO4 solution to which metallic copper is added; evaluation of IGC due to Cr23C6 formation is based on postexposure visual examination of specimens for fissures after bending. In Practice F, a 120 hour boiling test in CuSO4 + 50% H2SO4 solution (with metallic copper present), weight-loss corrosion rates indicate the degree of sensitization due to Cr23C6 precipitation for molybdenum-bearing stainless steels. Similarly, tests for detecting susceptibility to IGC in ferritic stainless steels have been incorporated into ASTM A 763 (Ref 7) and for wrought Ni-rich, Cr-bearing alloys, into ASTM G 28 (Ref 8). Acceptance criteria for ferritic and austenitic stainless steels, high nickel-base alloys, and aluminum alloys are summarized in Table 1. Table 1 Appropriate evaluation tests and acceptance criteria for wrought alloys UNS No.

Alloy name

Sensitizing treatment

Applicable tests (ASTM standards)

Exposure time, h

S43000

Type 430

None

24

S44600

Type 446

None

72

0.25 (10)

S44625

26-1

None

120

S44626

26-1S

None

120

0.05 (2) and no significant grain dropping No significant grain dropping

S44700

29-4

None

Ferric sulfate (A 763-X) Ferric sulfate (A 763-X) Ferric sulfate (A 763-X) Cupric sulfate (A 763-Y) Ferric sulfate (A

Criteria for passing, appearance or maximum allowable corrosion rate, mm/month (mils/month) 1.14 (45)

120

No significant grain dropping

S44800

29-4-2

None

S30400

Type 304

None

S30403

Type 304L

1 h at 675 °C (1250 °F)

S30908

Type 309S

None

S31600

Type 316

None

S31603

S31700

S31703

Type 316L

Type 317

Type 317L

1 h at 675 °C (1250 °F)

None

1 h at 675 °C (1250 °F)

S32100

Type 321

1 h at 675 °C (1250 °F) 1 h at 675 °C (1250 °F) 1 h at 675 °C (1250 °F) None

S34700

Type 347

N08020

20Cb-3

N08904

904L

N08825

Incoloy 825

N06007

Hastelloy G

1 h at 675 °C (1250 °F) None

N06985

Hastelloy G-3

None

N06625

Inconel 625

None

N06690

Inconel 690

N10276

1 h at 540 °C (1000 °F) Hastelloy C-276 None

N06455

Hastelloy C-4

None

N06110

Allcorr

None

763-X) Ferric sulfate (A 763-X) Oxalic acid (A 262-A) Ferric sulfate (A 262-B) Oxalic acid (A 262-A) Nitric acid (A 262-C) Nitric acid (A 262-C) Oxalic acid (A 262-A) Ferric sulfate (A 262-B) Oxalic acid (A 262-A) Ferric sulfate (A 262-B) Oxalic acid (A 262-A) Ferric sulfate (A 262-B) Oxalic acid (A 262-A) Ferric sulfate (A 262-B) Nitric acid (A262-C) Nitric acid (A 262-C) Ferric sulfate (G 28-A) Ferric sulfate (G 28-A) Nitric acid (A 262-C) Ferric sulfate (G 28-A) Ferric sulfate (G 28-A) Ferric sulfate (G 28-A) Nitric acid (A 262-C) Ferric sulfate (G 28-A) Ferric sulfate (G 28-A) Ferric sulfate (G 28-B)

120

No significant grain dropping

…

(a)

120

0.1 (4)

…

(a)

240

0.05 (2)

240

0.025 (1)

…

(a)

120

0.1 (4)

…

(a)

120

0.1 (4)

…

(a)

120

0.1 (4)

…

(a)

120

0.1 (4)

240

0.05 (2)

240

0.05 (2)

120

0.05 (2)

120

0.05 (2)

240

0.075 (3)

120

120

0.043 (1.7) sheet, plate, and bar; 0.05 (2) pipe and tubing 0.043 (1.7) sheet, plate, and bar; 0.05 (2) pipe and tubing 0.075 (3)

240

0.025 (1)

24

1 (40)

24

0.43 (17)

24

0.64 (25)

120

N10001

Hastelloy B

None

N10665

Hastelloy B-2

None

20% hydrochloric acid 20% hydrochloric acid Concentrated nitric acid (G 67)

24 24

0.075 (3) sheet, plate, and bar; 0.1 (4) pipe and tubing 0.05 (2) sheet, plate, and bar; 0.086 (3.4) pipe and tubing

(b) None 24 Aluminum Association 5xxx alloys (a) See ASTM A 262, practice A. (b) See ASTM G 67, section 4.1. Because sensitized alloys may inadvertently be used, acceptance tests are implemented as a quality-control check to evaluate stainless steels and nickel-base alloys when:

A95005– A95657

• • •

Different alloys, or alloys with “high” carbon content, are substituted for the low-carbon grades (for example, Type 316 substituted for Type 316L), and when welding or heat treatment are involved An improper heat treatment during fabrication results in the formation of intermetallic phases. The specified limits for carbon and/or nitrogen contents of an alloy are inadvertently exceeded.

Some standard tests include acceptance criteria, but others do not (Ref 4, 5). Suitable criteria are needed that can clearly separate material susceptible to IGA from that resistant to attack. Table 1 lists evaluation tests and acceptance criteria for various stainless steels and nickel-base alloys. Despite establishment of “standard” acceptance/rejection criteria, the buyer and seller can agree on a different criterion that meets their particular needs. ASTM G 108 describes a laboratory procedure for conducting a nondestructive electrochemical reactivation (EPR) test on Types 304 and 304L stainless steel to quantify the degree of sensitization (Ref 9). The metallographically mounted and highly polished test specimen is potentiodynamically polarized from the normally passive condition, in 0.5 M H2SO4 + 0.01 M KSCN solution at 30 ± 1 °C, to active potentials—a process known as reactivation. The amount of charge passed is related to the degree of IGC associated with Cr23C6 precipitation, which occurs predominately at the grain boundaries. After the single loop EPR test, the microstructure is examined and: 1. The grain boundary area is calculated from the grain size and the total exposed area of the test specimen. 2. Relative proportions of grain boundary attack and non-grain-boundary pitting are determined. A charge per unit grain-boundary area of 0.4 C/cm2 (0.062 C/in.2) indicates a heavily sensitized alloy. Although a double loop EPR method has been proposed (Ref 10, 11) to eliminate non-grain- boundary pitting and surface-finish effects, often observed in the single loop method, the double- loop technique is not presently included in ASTM G 108.

References cited in this section 1. M. Henthorne, Localized Corrosion—Cause of Metal Failure, STP 516, ASTM International, 1972, p 66–119 2. M.A. Streicher, Intergranular Corrosion of Stainless Alloys, STP 656, ASTM International, 1978, p 3– 84 3. M.A. Streicher, Intergranular Corrosion, Corrosion Tests and Standards: Application and Interpretation, R. Baboian, Ed., ASTM International, 1995, p 197–217 4. W.H. Hatfield, J. Iron Steel Inst., Vol 127, 1933, p 380–383 5. W.R. Huey, Trans. Am. Soc. Steel Treat., Vol 18, 1930, p 1126–1143

6. “Practices for Detecting Susceptibility to Intergranular Corrosion in Austenitic Stainless Steels,” A 262, Steel—Plate, Sheet, Strip, Wire; Stainless Steel Bar, Annual Book of ASTM Standards 2002, Vol 1.03, ASTM International, 2002 7. “Practices for Detecting Susceptibility to Intergranular Corrosion in Ferritic Stainless Steels”, A 763, Steel—Plate, Sheet, Strip, Wire; Stainless Steel Bar, Annual Book of ASTM Standards 2002, Vol 1.03, ASTM International, 2002 8. “Standard Test Methods for Detecting Susceptibility to Intergranular Corrosion in Wrought, NickelRich, Chromium-Bearing Alloys,” G 28, Wear and Erosion; Metal Corrosion, Annual Book of ASTM Standards 2002, Vol 3.02, ASTM International, 2002 9. “Standard Test Method for Electrochemical Reactivation (EPR) for Detecting Sensitization of AISI Type 304 and 304L Stainless Steel,” G 108, Wear and Erosion; Metal Corrosion, Annual Book of ASTM Standards 2002, Vol 3.02, ASTM International, 2002 10. A.P. Majidi and M.A. Striecher, Corrosion, Vol 40, 1984, p 584–592 11. M. Akashi, T. Kawamoto, and F. Umemura, Corros. Eng., (Jpn.), Vol 29, 1980, p 163


Tests for Aluminum Alloys The electrochemically active paths at the grain boundaries of aluminum alloy materials can be either the solid solution or closely spaced anodic second-phase intermetallic particles. The identities of the specific actively corroding paths vary with the alloy composition and metallurgical condition of the product, as discussed in Ref 12 and 13. The most serious forms of such structure-dependent corrosion are stress-corrosion cracking (SCC) and exfoliation. Stress-corrosion cracking requires the presence of a sustained (residual and/or service) tensile stress, and exfoliation occurs only in wrought products with a directional grain structure. Not all materials that are susceptible to IGA, however, are susceptible to SCC or exfoliation. Strain-Hardened 5xxx (Al-Mg) Alloys. Alloys in this series that contain more than approximately 3% Mg are rendered susceptible to IGA (sensitization) by certain manufacturing conditions or by being subjected to elevated temperatures up to approximately 175 °C (350 °F). This is the result of the continuous grain-boundary precipitation of the highly anodic Mg2Al3 phase, which corrodes preferentially in most corrosive environments. ASTM G 67 is a method that provides a quantitative measure of the susceptibility to IGA of these alloys (Ref 14). This method consists of immersing test specimens in concentrated HNO3 at 30 °C (85 °F) for 24 h and determining the mass loss per unit area of exposed surfaces as the measure of intergranular susceptibility. When the second phase is precipitated in a relatively continuous network along grain boundaries, the preferential attack of the network causes whole grains to drop out of the specimens. Such dropping out causes relatively large mass losses of the order of 25 to 75 mg/cm2 (3.9 - 12 mg/in.2), whereas specimens of materials resistant to IGC lose only about 1 to 15 mg/cm2 (0.2 - 2.3 mg/in.2). Intermediate mass losses occur when the precipitate is randomly distributed. The parallel relationship between the susceptibility to intergranular attack, SCC, and exfoliation of these particular alloys makes ASTM G 67 a useful screening test for alloy and process

development studies. A problem arises, however, in selecting a pass-or-fail value in relation to the performance of intermediate materials in environments other than HNO3. Heat Treated High-Strength Alloys. Materials problems caused by SCC, exfoliation, or corrosion fatigue of the early 2xxx (aluminum- copper) alloys were related to IGC, and the blame came to be associated with improper heat treatment. In 1944, an accelerated test for detecting susceptibility to IGC was incorporated into a U.S. Government specification for the heat treatment of aluminum alloys. Military Specification MIL-H-6088F has superseded this specification. Tests are required for periodic monitoring of 2xxx and 7xxx series alloys in all rivets and fastener components as well as sheet, bar, rod, wire, and shapes under 6.4 mm (0.25 in.) thick. Specimen preparation, test procedure, and evaluation criteria are detailed in Ref 15. Other Tests for Aluminum Alloys. The volume of hydrogen evolved on immersion of etched 2xxx series (aluminum-copper-magnesium) alloys in a solution containing 3% sodium chloride (NaCl) and 1% hydrochloric acid (HCl) for a stipulated time has been used as a quantitative measure of the severity of IGA. A problem with this approach is that the correlation between the amount (or the rate) of hydrogen evolved is influenced by a number of factors, including alloy composition, temper, and grain size (Ref 16, 17). Applied current or potential in neutral chloride solutions (for example, 0.1 N NaCl) provides another direct method of assessing the degree of susceptibility to intergranular attack when accompanied by a microscopic examination of metallographic sections (Ref 16, 18, 19). As stated earlier, exfoliation is a form of corrosion that can occur in layers parallel to rolled surfaces, especially on aluminum alloys, and attack is generally along grain boundaries. ASTM G 34 is an accelerated test method to determine exfoliation corrosion susceptibility of 2xxx and 7xxx aluminum alloys (Ref 20); this is sometimes known as the EXCO test. Specimens of wrought material are immersed in 4 M NaCl + 0.5 M KNO3 + 0.1 M HNO3 solution at 25 ± 3 °C. Immersion times are 96 and 48 h for 2xxx and 7xxx alloys, respectively. Performance ratings are established by reference to standard photographs. An analogous exfoliation test for wrought 5xxx aluminum alloys, containing ≥2% Mg, is covered by ASTM G 66 in which test specimens are immersed in 1 M NH4Cl + 0.25 M NH4NO3 + 0.01 M (NH4)2C4H4O6 + 0.09 M H2O2 for 24 h at 65 ± 1 °C, followed by visual comparison with standard photographs (Ref 21). This test is sometimes known as “ASSET,” an abbreviation for assessment of exfoliation corrosion test.

References cited in this section 12. T.J. Summerson and D.O. Sprowls, Corrosion Behavior of Aluminum Alloys, Aluminum Alloys: Their Physical and Mechanical Properties, Vol III, E.A. Starke, Jr. and T.H. Sanders, Jr., Ed., Engineering Materials Advisory Services Ltd., 1986, p 1576–1662 13. B.W. Lifka and D.O. Sprowls, Localized Corrosion—Cause of Metal Failure, STP 516, ASTM International, 1972, p 120–144 14. “Standard Test Method for Detecting Susceptibility to Intergranular Corrosion of 5xxx Series Aluminum Alloys by Mass Loss After Exposure to Nitric Acid (NAMLT Test)”, G 67, Wear and Erosion; Metal Corrosion, Annual Book of ASTM Standards 2002, Vol 3.02, ASTM International, 2002 15. “Heat Treatment of Aluminum Alloys,” Military Specification MIL-H-6088F, United States Government Printing Office 16. F.A. Champion, Corrosion Testing Procedures, 2nd ed., John Wiley & Sons, 1965, p 365, 366 17. G.J. Schafer, J. Appl. Chem., Vol 10, 1960, p 138 18. S. Ketcham and W. Beck, Corrosion, Vol 16, 1960, p 37 19. M.K. Budd and F.F. Booth, Corrosion, Vol 18, 1962, p 197

20. “Standard Test Method for Exfoliation Corrosion Susceptibility in 2xxx and 7xxx Series Aluminum Alloys (EXCO Test)”, G 34, Wear and Erosion; Metal Corrosion, Annual Book of ASTM Standards 2002, Vol 3.02, ASTM International, 2002 21. “Standard Test Method for Visual Assessment of Exfoliation Corrosion Susceptibility in 5xxx Series Aluminum Alloys (ASSET Test)”, G 66, Wear and Erosion; Metal Corrosion, Annual Book of ASTM Standards 2002, Vol 3.02, ASTM International, 2002


Tests for Other Alloys Although IGC is present to some extent in metallic materials other than stainless steels and aluminum alloys, incidences of attack associated with this form of corrosion are few and are generally not of practical importance. Therefore, no attempts have been made to develop and standardize specific tests for detecting susceptibility to IGC in these alloys. However, certain media have been commonly used for evaluating the susceptibility to IGC of magnesium, copper, lead, and zinc alloys (Ref 16). These media are listed in Table 2. The presence or absence of attack in these tests is not necessarily a measure of the performance of the material in any other corrosive (test or service) environments. Table 2 Media for testing susceptibility to intergranular corrosion Alloy

Medium

Concentration %

Magnesium alloys Copper alloys

Sodium chloride plus hydrochloric acid

…

Lead alloys Zinc alloys

Temperature (°F) Room

°C

Sodium chloride plus sulfuric or nitric 1 NaCl, 0.3 acid 40–50 (105–120) acid Acetic acid or hydrochloric acid … Room Humid air 100% relative 95 (205) humidity

Source: Ref 16 Magnesium Alloys. There are rare instances of reported IGC of magnesium alloys, as in the case of chromic acid contaminated with chlorides or sulfates. The copper alloys that appear to be the most susceptible to IGC are Muntz metal, admiralty brass, aluminum brasses, and silicon bronzes. Admiralty alloys have been observed to suffer IGC on exposure to saline cooling waters, although the incidence of attack is very low. The antimony-doped grades are reportedly superior to the arsenical grades in this respect. Similarly, arsenical and phosphorized grades of aluminum brass have been observed to suffer IGC in seawater-type exposures. Zinc die casting alloys have reportedly suffered IGC in certain steam atmospheres. A laboratory test for simulating service failures, and particularly for alloy development work, has been in use for testing the susceptibility of zinc- base die-castings to IGC (Ref 22). The test consists of exposing samples to air at 95 °C (205 °F) and 100% relative humidity for 10 days under conditions permitting condensation of hot water on the metal. Susceptibility to IGC is assessed by the effect on mechanical properties, such as impact strength.

Experience has shown that castings with mechanical properties and dimensions that are not significantly altered by the 10 day exposure in this test will not suffer IGA in atmospheric service. Intergranular corrosion at elevated temperatures can occur by formation of low melting phases (in certain alloyenvironment combinations) that result in rapid attack commonly at grain boundaries. For example, the formation of nickel-sulfides in nickel-base alloys can cause catastrophic failures in high-temperature, sulfurrich gaseous environments. This phenomenon is commonly referred to as sulfidation. Deep penetration can occur rapidly, including through the full thickness of the metal. The techniques for evaluating this type of IGC include (a) x-ray mapping during examination in a scanning electron microscope (equipped with an energydispersive x-ray detector) and (b) transmission electron microscopy.

References cited in this section 16. F.A. Champion, Corrosion Testing Procedures, 2nd ed., John Wiley & Sons, 1965, p 365, 366 22. H.H. Uhlig, Corrosion Handbook, John Wiley & Sons, 1948


Acknowledgment This article has been adapted from Donald O. Sprowls, Evaluation of Intergranular Corrosion, Corrosion, Volume 13, ASM Handbook (formerly 9th ed. Metals Handbook), ASM International, 1987, pages 239 to 241.


References 1. M. Henthorne, Localized Corrosion—Cause of Metal Failure, STP 516, ASTM International, 1972, p 66–119 2. M.A. Streicher, Intergranular Corrosion of Stainless Alloys, STP 656, ASTM International, 1978, p 3– 84 3. M.A. Streicher, Intergranular Corrosion, Corrosion Tests and Standards: Application and Interpretation, R. Baboian, Ed., ASTM International, 1995, p 197–217 4. W.H. Hatfield, J. Iron Steel Inst., Vol 127, 1933, p 380–383

5. W.R. Huey, Trans. Am. Soc. Steel Treat., Vol 18, 1930, p 1126–1143 6. “Practices for Detecting Susceptibility to Intergranular Corrosion in Austenitic Stainless Steels,” A 262, Steel—Plate, Sheet, Strip, Wire; Stainless Steel Bar, Annual Book of ASTM Standards 2002, Vol 1.03, ASTM International, 2002 7. “Practices for Detecting Susceptibility to Intergranular Corrosion in Ferritic Stainless Steels”, A 763, Steel—Plate, Sheet, Strip, Wire; Stainless Steel Bar, Annual Book of ASTM Standards 2002, Vol 1.03, ASTM International, 2002 8. “Standard Test Methods for Detecting Susceptibility to Intergranular Corrosion in Wrought, NickelRich, Chromium-Bearing Alloys,” G 28, Wear and Erosion; Metal Corrosion, Annual Book of ASTM Standards 2002, Vol 3.02, ASTM International, 2002 9. “Standard Test Method for Electrochemical Reactivation (EPR) for Detecting Sensitization of AISI Type 304 and 304L Stainless Steel,” G 108, Wear and Erosion; Metal Corrosion, Annual Book of ASTM Standards 2002, Vol 3.02, ASTM International, 2002 10. A.P. Majidi and M.A. Striecher, Corrosion, Vol 40, 1984, p 584–592 11. M. Akashi, T. Kawamoto, and F. Umemura, Corros. Eng., (Jpn.), Vol 29, 1980, p 163 12. T.J. Summerson and D.O. Sprowls, Corrosion Behavior of Aluminum Alloys, Aluminum Alloys: Their Physical and Mechanical Properties, Vol III, E.A. Starke, Jr. and T.H. Sanders, Jr., Ed., Engineering Materials Advisory Services Ltd., 1986, p 1576–1662 13. B.W. Lifka and D.O. Sprowls, Localized Corrosion—Cause of Metal Failure, STP 516, ASTM International, 1972, p 120–144 14. “Standard Test Method for Detecting Susceptibility to Intergranular Corrosion of 5xxx Series Aluminum Alloys by Mass Loss After Exposure to Nitric Acid (NAMLT Test)”, G 67, Wear and Erosion; Metal Corrosion, Annual Book of ASTM Standards 2002, Vol 3.02, ASTM International, 2002 15. “Heat Treatment of Aluminum Alloys,” Military Specification MIL-H-6088F, United States Government Printing Office 16. F.A. Champion, Corrosion Testing Procedures, 2nd ed., John Wiley & Sons, 1965, p 365, 366 17. G.J. Schafer, J. Appl. Chem., Vol 10, 1960, p 138 18. S. Ketcham and W. Beck, Corrosion, Vol 16, 1960, p 37 19. M.K. Budd and F.F. Booth, Corrosion, Vol 18, 1962, p 197 20. “Standard Test Method for Exfoliation Corrosion Susceptibility in 2xxx and 7xxx Series Aluminum Alloys (EXCO Test)”, G 34, Wear and Erosion; Metal Corrosion, Annual Book of ASTM Standards 2002, Vol 3.02, ASTM International, 2002 21. “Standard Test Method for Visual Assessment of Exfoliation Corrosion Susceptibility in 5xxx Series Aluminum Alloys (ASSET Test)”, G 66, Wear and Erosion; Metal Corrosion, Annual Book of ASTM Standards 2002, Vol 3.02, ASTM International, 2002 22. H.H. Uhlig, Corrosion Handbook, John Wiley & Sons, 1948



A.H. Ailor, Ed., Handbook of Corrosion Testing and Evaluation, J. Wiley & Sons, Inc., 1971 R. Baboian, Ed., Manual 20 Corrosion Tests and Standards: Application and Interpretation, ASTM International, 1995 M. Fontana, Corrosion Engineering, 3rd ed., McGraw-Hill, 1986 High Temperature Corrosion, NACE International, 1983 G.Y. Lai, High Temperature Corrosion of Engineering Alloys, ASM International, 1990

B. Phull, Evaluating Exfoliation Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 572–574

Evaluating Exfoliation Corrosion Revised by Bopinder Phull, Consultant

Introduction EXFOLIATION is a structure-dependent form of localized (usually) intergranular corrosion that is most idiosyncratic in certain alloys and tempers of aluminum. It is also called layer corrosion or lamellar corrosion and is a form of subsurface attack that follows narrow paths (often grain boundaries) in the rolling direction of wrought materials and parallel to the rolled surfaces. Exfoliation corrosion often initiates at laterally sheared edges exposed to the environment. The corrosion products tend to swell or pry open the material, revealing alternate layers or leafing of corroded and uncorroded structure. Highly cold-worked materials, which have elongated grain boundaries, and thin-gage product forms tend to be most affected. Swelling associated with exfoliation corrosion can stress adjacent parts and cause overload failures—a phenomenon akin to packout rusting (Ref 1). The occurrence of exfoliation in susceptible materials is influenced to a marked degree by environmental conditions. Figure 1 illustrates the broad range of behavior of aluminum alloy 2124 plate in different types of environments (Ref 2). The plate was heat treated to be susceptible to exfoliation. Performance can vary significantly, depending on the environment. For example, forged truck wheels made of an aluminum-copper alloy (2024-T4) give corrosion- free service for many years in the warm climates of the southern and western United States, but they exfoliate severely in only 1 or 2 years in the northern states, where deicing salts are used on the highways during the winter months.

Fig. 1 Comparison of exfoliation of aluminum alloy 2124 (heat treated to be susceptible; EXCO ED rating) in various seacoast and industrial environments. Specimens were 13 mm ( in.) plate. Source: Ref 2 Accelerated laboratory tests do not precisely predict long-term corrosion behavior; however, answers are needed quickly in the development of new materials. For this reason, accelerated tests are used to screen candidate alloys before conducting atmospheric exposures or other field tests. They are also sometimes used for quality- control tests. Several new laboratory tests for exfoliation corrosion have been standardized under the jurisdiction of American Society for Testing and Materials International (ASTM International) Committee G-1 on the Corrosion of Metals.

References cited in this section 1. J.B. Vrable, R.T. Jones, and E.H. Phelps, “The Application of High Strength Low Alloy Steels in the Chemical Industry,” Mater. Perform., Vol 18 (No. 1), Jan 1979, p 39–44 2. S.J. Ketcham and E.J. Jankowsky, Developing an Accelerated Test: Problems and Pitfalls. Laboratory Corrosion Tests and Standards, STP 866, G.S. Haynes and R. Baboian, Ed., American Society for Testing and Materials, 1985, p 14–23

B. Phull, Evaluating Exfoliation Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 572–574 Evaluating Exfoliation Corrosion Revised by Bopinder Phull, Consultant

Spray Tests

Three cyclic acidified salt spray tests have been widely used in the aluminum and aircraft industries. These are covered by the procedures described in Annexes A2, A3, and A4 of ASTM G 85 (Ref 3). This standard does not prescribe the particular practice, test specimen, or exposure period to be used for a specific product, nor does it define the interpretation to be given to the test results. These considerations are prescribed by specifications covering the material or product being tested or by agreement between the purchaser and the seller. Annex A2 describes a cyclic salt spray test that uses a 5% sodium chloride (NaCl) solution, acidified to pH 3 with acetic acid, in a spray chamber at a temperature of 49 ± 1 °C (120 ± 2 °F). This test is applicable for exfoliation testing of 2xxx (dry-bottom operation) and 7xxx (wet-bottom operation; that is, with approximately 25 mm, or 1 in., of water present in the bottom of the test chamber) aluminum alloys with a test duration of 1 to 2 weeks. Results with 7075 and 7178 alloys in various metallurgical conditions have been shown to correlate well with results obtained in a seacoast atmosphere (4 year exposure at Point Judith, RI) (Ref 4). Annex A3 describes another cyclic salt spray test that uses a 5% synthetic sea salt solution, acidified to pH 3 with acetic acid, in a spray chamber at a temperature of 49 °C (120 °F). The test is applicable to the production control of exfoliation-resistant tempers of the 2xxx, 5xxx, and 7xxx aluminum alloys (Ref 5, 6). Wet-bottom operating conditions are recommended with test durations of 1 to 2 weeks. Annex A4 describes a salt/sulfur dioxide (SO2) spray test that uses either 5% NaCl or 5% synthetic sea salt solution in a spray chamber at a temperature of 35 °C (95 °F). The spray may be either cyclic or constant. This, along with the type of salt solution and the test duration, is subject to agreement between the purchaser and the seller. The test is applicable for 2xxx and 7xxx aluminum alloys. Test duration is 2 to 4 weeks (Ref 2).

References cited in this section 2. S.J. Ketcham and E.J. Jankowsky, Developing an Accelerated Test: Problems and Pitfalls. Laboratory Corrosion Tests and Standards, STP 866, G.S. Haynes and R. Baboian, Ed., American Society for Testing and Materials, 1985, p 14–23 3. “Standard Practice for Modified Salt Spray (Fog) Testing,” G 85, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards 2002, American Society for Testing and Materials, 2002 4. B.W. Lifka and D.O. Sprowls, Relationship of Accelerated Test Methods for Exfoliation Resistance in 7xxx Aluminum Alloys with Exposure to a Seacoast Atmosphere. Corrosion in Natural Environments, STP 558, American Society for Testing and Materials, 1974, p 306–333 5. H.B. Romans, An Accelerated Laboratory Test to Determine the Exfoliation Corrosion Resistance of Aluminum Alloys, Mater. Res. Stand., Vol 9 (No. 11), 1969, p 31–34 6. S.J. Ketcham and P.W. Jeffrey, Exfoliation Corrosion Testing of 7178 and 7075 Aluminum Alloys, Localized Corrosion— Cause of Metal Failure, STP 516, American Society for Testing and Materials, 1972, p 273–302


Immersion Tests

Total-immersion tests were developed to provide simpler, more easily controlled test methods. Chloride solutions did not cause exfoliation during reasonable periods of immersion. However, formulations of chloridenitrate solutions were found that produced severe exfoliation of susceptible alloys in only 1 or 2 days. Optimal test conditions differed for separate alloy families (Ref 7). ASTM G 66 describes a procedure for the continuous-immersion exfoliation testing of 5xxx alloys containing 2.0% or more magnesium (Ref 8). Specimens are immersed for 24 h at 65 ± 1 °C (150 ± 2 °F) in a solution of 1 M ammonium chloride, 0.25 M ammonium nitrate, 0.01 M ammonium tartrate, and 0.09 M hydrogen peroxide. Susceptibility to exfoliation is determined by visual examination, using performance ratings established by reference to standard photographs. This method is stated to provide reliable prediction of the exfoliation corrosion behavior of 5xxx alloys in marine environments (Ref 9). The test is also useful for alloy development studies and quality control of mill products such as sheet and plate (Ref 10). ASTM G 34 provides an accelerated exfoliation corrosion test for 2xxx and 7xxx aluminum alloys through the continuous immersion of test materials in an aqueous solution of 4 M NaCl, 0.5 M potassium nitrate, and 0.1 M nitric acid at 25 ± 3 °C (77 ± 5 °F) (Ref 11). Maximum recommended exposure periods for the 2xxx and 7xxx aluminum alloys are 96 and 48 h, respectively. Susceptibility to exfoliation is determined by visual examination, using performance ratings established by reference to standard photographs. This constant immersion exfoliation corrosion test method, also known as the EXCO test, is primarily used for research and development and quality control of such mill products as sheet and plate (Ref 10). However, it should not be construed as the optimal method for quality acceptance. The ASTM G 34 method provides a useful prediction of the exfoliation behavior of 2xxx and 7xxx aluminum alloys in various types of outdoor service, especially in marine and industrial environments (Ref 4, 12). The test solution is very corrosive and represents the more severe types of environment service (Fig. 1). However, it remains to be determined whether correlations can be established between EXCO test ratings and practical service conditions for a given alloy. It has been reported that samples of 7xxx (Al-Zn- Mg-Cu) alloys rated EA (superficial exfoliation) or P (pitting) in a 48 h EXCO test did not develop more than superficial exfoliation (EA rating) during 6 to 9 year exposures to seacoast atmospheres, while EC- and ED-rated (severe and very severe exfoliation, respectively) materials developed severe exfoliation within 1 to 7 years at the seacoast (specimens rated EA to ED are shown in Fig. 2, Fig. 3, Fig. 4, Fig. 5) (Ref 12).

Fig. 2 Examples of exfoliation rating EA (superficial). Specimens exhibit tiny blisters, thin slivers, flakes, or powder, with only slight separation of metal. Source: Ref 11

Fig. 3 Examples of exfoliation rating EB (moderate). Specimens show notable layering and penetration into the metal. Source: Ref 11

Fig. 4 Examples of exfoliation rating EC (severe). There is penetration to a considerable depth into the metal. Source: Ref 11

Fig. 5 Examples of exfoliation rating ED (very severe). Specimens appear similar to EC except for much greater penetration and loss of metal. Source: Ref 11 Performance differences between practical service and the EXCO test have been noted and indicate that the EXCO test may be too severe for some of the more recently developed 2xxx and 7xxx alloys. The testing program for evaluating new alloy materials should consist of multiple tests with one of the less aggressive ASTM G 85 salt spray methods supplemented by outdoor exposure tests. Caution must be exercised in setting limits for material procurement specifications based on accelerated tests (Ref 13).

References cited in this section 4. B.W. Lifka and D.O. Sprowls, Relationship of Accelerated Test Methods for Exfoliation Resistance in 7xxx Aluminum Alloys with Exposure to a Seacoast Atmosphere. Corrosion in Natural Environments, STP 558, American Society for Testing and Materials, 1974, p 306–333 7. D.O. Sprowls, J.D. Walsh, and M.B. Shumaker, Simplified Exfoliation Testing of Aluminum Alloys, Localized Corrosion— Cause of Metal Failure, STP 516, American Society for Testing and Materials, 1972, p 38–65 8. “Standard Test Method for Visual Assessment of Exfoliation Corrosion Susceptibility of 5xxx-Series Aluminum Alloys (ASSET Test),” G 66, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards 2002, American Society for Testing and Materials, 2002 9. T.J. Summerson, Interim Report, Aluminum Association Task Group on Exfoliation and StressCorrosion Cracking of Aluminum Alloys for Boat Stock, Proceedings of the Tri-Service Corrosion Military Equipment Conference, Technical Report AFML-TR- 75-42, Vol II, Air Force Materials Laboratory, 1975, p 193–221 10. “Specification for Aluminum and Aluminum-Alloy Sheet and Plate,” B 209, Aluminum and Magnesium Alloys, Vol 02.02, Annual Book of ASTM Standards 2002, American Society for Testing and Materials, 2002

11. “Standard Test Method for Exfoliation Corrosion Susceptibility in 2xxx and 7xxx-Series Aluminum Alloys (EXCO Test),” G 34, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards 2002, American Society for Testing and Materials, 2002 12. D.O. Sprowls, T.J. Summerson, and F.E. Loftin, Exfoliation Corrosion Testing of 7075 and 7178 Aluminum Alloys—Interim Report on Atmospheric Exposure Tests (Report of ASTM G01.05.02 Interlaboratory Testing Program in Cooperation with the Aluminum Association), Corrosion in Natural Environments, STP 558, American Society for Testing and Materials, 1974, p 99–113 13. B.W. Lifka, “Corrosion Resistance of Aluminum Alloy Plate in Rural, Industrial, and Seacoast Atmospheres,” Paper 420, Corrosion/87, NACE International, 1987


Acknowledgment This article has been adapted from D.O. Sprowls, Evaluation of Exfoliation Corrosion, Corrosion, Vol 13, ASM Handbook (formerly Metals Handbook, 9th ed.), ASM International, 1987, p 242–244


Acknowledgment This article has been adapted from D.O. Sprowls, Evaluation of Exfoliation Corrosion, Corrosion, Vol 13, ASM Handbook (formerly Metals Handbook, 9th ed.), ASM International, 1987, p 242–244


References

1. J.B. Vrable, R.T. Jones, and E.H. Phelps, “The Application of High Strength Low Alloy Steels in the Chemical Industry,” Mater. Perform., Vol 18 (No. 1), Jan 1979, p 39–44 2. S.J. Ketcham and E.J. Jankowsky, Developing an Accelerated Test: Problems and Pitfalls. Laboratory Corrosion Tests and Standards, STP 866, G.S. Haynes and R. Baboian, Ed., American Society for Testing and Materials, 1985, p 14–23 3. “Standard Practice for Modified Salt Spray (Fog) Testing,” G 85, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards 2002, American Society for Testing and Materials, 2002 4. B.W. Lifka and D.O. Sprowls, Relationship of Accelerated Test Methods for Exfoliation Resistance in 7xxx Aluminum Alloys with Exposure to a Seacoast Atmosphere. Corrosion in Natural Environments, STP 558, American Society for Testing and Materials, 1974, p 306–333 5. H.B. Romans, An Accelerated Laboratory Test to Determine the Exfoliation Corrosion Resistance of Aluminum Alloys, Mater. Res. Stand., Vol 9 (No. 11), 1969, p 31–34 6. S.J. Ketcham and P.W. Jeffrey, Exfoliation Corrosion Testing of 7178 and 7075 Aluminum Alloys, Localized Corrosion— Cause of Metal Failure, STP 516, American Society for Testing and Materials, 1972, p 273–302 7. D.O. Sprowls, J.D. Walsh, and M.B. Shumaker, Simplified Exfoliation Testing of Aluminum Alloys, Localized Corrosion— Cause of Metal Failure, STP 516, American Society for Testing and Materials, 1972, p 38–65 8. “Standard Test Method for Visual Assessment of Exfoliation Corrosion Susceptibility of 5xxx-Series Aluminum Alloys (ASSET Test),” G 66, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards 2002, American Society for Testing and Materials, 2002 9. T.J. Summerson, Interim Report, Aluminum Association Task Group on Exfoliation and StressCorrosion Cracking of Aluminum Alloys for Boat Stock, Proceedings of the Tri-Service Corrosion Military Equipment Conference, Technical Report AFML-TR- 75-42, Vol II, Air Force Materials Laboratory, 1975, p 193–221 10. “Specification for Aluminum and Aluminum-Alloy Sheet and Plate,” B 209, Aluminum and Magnesium Alloys, Vol 02.02, Annual Book of ASTM Standards 2002, American Society for Testing and Materials, 2002 11. “Standard Test Method for Exfoliation Corrosion Susceptibility in 2xxx and 7xxx-Series Aluminum Alloys (EXCO Test),” G 34, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards 2002, American Society for Testing and Materials, 2002 12. D.O. Sprowls, T.J. Summerson, and F.E. Loftin, Exfoliation Corrosion Testing of 7075 and 7178 Aluminum Alloys—Interim Report on Atmospheric Exposure Tests (Report of ASTM G01.05.02 Interlaboratory Testing Program in Cooperation with the Aluminum Association), Corrosion in Natural Environments, STP 558, American Society for Testing and Materials, 1974, p 99–113 13. B.W. Lifka, “Corrosion Resistance of Aluminum Alloy Plate in Rural, Industrial, and Seacoast Atmospheres,” Paper 420, Corrosion/87, NACE International, 1987


References 1. J.B. Vrable, R.T. Jones, and E.H. Phelps, “The Application of High Strength Low Alloy Steels in the Chemical Industry,” Mater. Perform., Vol 18 (No. 1), Jan 1979, p 39–44 2. S.J. Ketcham and E.J. Jankowsky, Developing an Accelerated Test: Problems and Pitfalls. Laboratory Corrosion Tests and Standards, STP 866, G.S. Haynes and R. Baboian, Ed., American Society for Testing and Materials, 1985, p 14–23 3. “Standard Practice for Modified Salt Spray (Fog) Testing,” G 85, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards 2002, American Society for Testing and Materials, 2002 4. B.W. Lifka and D.O. Sprowls, Relationship of Accelerated Test Methods for Exfoliation Resistance in 7xxx Aluminum Alloys with Exposure to a Seacoast Atmosphere. Corrosion in Natural Environments, STP 558, American Society for Testing and Materials, 1974, p 306–333 5. H.B. Romans, An Accelerated Laboratory Test to Determine the Exfoliation Corrosion Resistance of Aluminum Alloys, Mater. Res. Stand., Vol 9 (No. 11), 1969, p 31–34 6. S.J. Ketcham and P.W. Jeffrey, Exfoliation Corrosion Testing of 7178 and 7075 Aluminum Alloys, Localized Corrosion— Cause of Metal Failure, STP 516, American Society for Testing and Materials, 1972, p 273–302 7. D.O. Sprowls, J.D. Walsh, and M.B. Shumaker, Simplified Exfoliation Testing of Aluminum Alloys, Localized Corrosion— Cause of Metal Failure, STP 516, American Society for Testing and Materials, 1972, p 38–65 8. “Standard Test Method for Visual Assessment of Exfoliation Corrosion Susceptibility of 5xxx-Series Aluminum Alloys (ASSET Test),” G 66, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards 2002, American Society for Testing and Materials, 2002 9. T.J. Summerson, Interim Report, Aluminum Association Task Group on Exfoliation and StressCorrosion Cracking of Aluminum Alloys for Boat Stock, Proceedings of the Tri-Service Corrosion Military Equipment Conference, Technical Report AFML-TR- 75-42, Vol II, Air Force Materials Laboratory, 1975, p 193–221 10. “Specification for Aluminum and Aluminum-Alloy Sheet and Plate,” B 209, Aluminum and Magnesium Alloys, Vol 02.02, Annual Book of ASTM Standards 2002, American Society for Testing and Materials, 2002 11. “Standard Test Method for Exfoliation Corrosion Susceptibility in 2xxx and 7xxx-Series Aluminum Alloys (EXCO Test),” G 34, Wear and Erosion; Metal Corrosion, Vol 03.02, Annual Book of ASTM Standards 2002, American Society for Testing and Materials, 2002

12. D.O. Sprowls, T.J. Summerson, and F.E. Loftin, Exfoliation Corrosion Testing of 7075 and 7178 Aluminum Alloys—Interim Report on Atmospheric Exposure Tests (Report of ASTM G01.05.02 Interlaboratory Testing Program in Cooperation with the Aluminum Association), Corrosion in Natural Environments, STP 558, American Society for Testing and Materials, 1974, p 99–113 13. B.W. Lifka, “Corrosion Resistance of Aluminum Alloy Plate in Rural, Industrial, and Seacoast Atmospheres,” Paper 420, Corrosion/87, NACE International, 1987

B. Phull, Evaluating Stress-Corrosion Cracking, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 575–616

Evaluating Stress-Corrosion Cracking Revised by Bopinder Phull, Consultant

Introduction THERE ARE A NUMBER of corrosion-related causes of the premature fracture of structural components. The most common of these are compared in Fig. 1. Cracking due to corrosion fatigue occurs only under cyclic or fluctuating operating loads, while failure resulting from the other processes shown occurs under static or slowly rising loads. With certain alloy systems, hydrogen embrittlement may have a contributory role in each of these failure processes. Appropriate tests for the different failure modes are discussed in other articles in this Section.

Fig. 1 Causes of premature fracture influenced by the corrosion of a structural component The materials presented here are organized according to the following broad outline: • • • • • • •

General state-of-the-art Static loading of smooth specimens Static loading of precracked specimens Dynamic loading: slow-strain-rate testing Selection of test environments Appropriate tests for various alloy systems Interpretation of test results

B. Phull, Evaluating Stress-Corrosion Cracking, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 575–616 Evaluating Stress-Corrosion Cracking Revised by Bopinder Phull, Consultant

Static Loading of Smooth Specimens Tests for predicting the stress-corrosion performance of an alloy in a particular service application should be conducted with a stress system similar to that anticipated in service. Table 1 lists the numerous sources of sustained tension that are known to have initiated SCC in service and the applicable methods of stressing. Most of the SCC service problems involve tensile stresses of unknown magnitude that are usually very high. Tests that incorporate a high total strain are usually the most realistic in terms of duplicating service. Table 1 Stressing methods applicable to various sources of sustained tension in service Source of sustained tension in service Constant strain Constant load Residual stress X … Quenching after heat treatment X … Forming X … Welding X … Misalignment (fit-up stresses) X … Interference fasteners Interference bushings X … Rigid … X Flexible X … Flareless fittings X … Clamps X X Hydraulic pressure … X Deadweight X X Faying surface corrosion Note: The greatest hazard arises when residual, assembly, and operating stresses are additive. The results are strongly influenced by the mechanical aspects of the tests, such as method of loading and specimen size. These mechanical aspects can have variable effects on the initiation and propagation lifetimes and can influence estimates of a threshold stress. Therefore, an apparent threshold stress for SCC is not a material property, and threshold estimates must be qualified with regard to the test conditions and the significance level.

Constant-Strain versus Constant-Load Tests Constant-strain (fixed-displacement) tests are widely used, primarily because a variety of simple and inexpensive stressing jigs can be devised. However, there is poor reproducibility of the exposure stress with some of these techniques. Therefore, sophisticated procedures have been developed to improve this facet of testing. Constant-strain tests are sometimes called decreasing-load tests, because after the onset of SCC in small test specimens the gross section exposure stress decreases. This results from the opening of the crack (or cracks) under the high stress concentration at the crack tip (or tips) and causes some of the applied elastic strain to change to plastic strain, with an attendant reduction in the initial load (Ref 6, 7). Such trends in changing stress during crack growth are shown in Fig. 4.

Fig. 4 Comparison of changing stress during initiation and growth of isolated SCC in constant-strain and constant-load tests of a uniaxially loaded tension specimen. (a) Constant-strain test. (b) Constantload test. σM is the maximum stress at crack tip, σN is the average stress in the net section, and σG is the applied stress to the gross section. Source: Ref 7 Comparison of the stress trends for a constant- strain test (Fig. 4a) with those for a constant-load test (Fig. 4b) reveals that neither method of loading provides a constant-stress test after growth of microcracks has occurred. True constant-load (dead-load) tests result in increasing stress levels as cracking progresses and are more likely to lead to earlier failure with complete fracture and lower estimates of a threshold stress than constant-strain tests. Figures 4(a) and (b) illustrate basic trends that may be applied to all types of test specimens, including precracked specimens. Specific curves, however, will differ depending on other test conditions. The stiffness of the combined stressing frame/ test specimen system can have a significant effect on materials evaluation if identical test procedures are not used (Ref 6). Many so-called constant-strain tests are not actually constant- strain tests, particularly if a spring is included in the stressing system, because a significant amount of elastic strain energy may be contained in the stressing system. Depending on the “softness” of the spring or the elasticity of the stressing jig, the stiffness (compliance) of the stressing system can be varied greatly between zero stiffness (dead load) and infinite stiffness (true constant total strain). Figure 5 shows the typical change in net section stress with the onset of SCC in an intermediate-stiffness stressing frame used in ASTM G 49 for loading direct-tension stress-corrosion test specimens (Ref 8).

Fig. 5 Effect of loading method and extent of cracking or corrosion pattern on average net section stress in a uniaxially loaded tension specimen. Behavior is generally representative, but curves will vary with specific alloys and tempers. (a) Localized cracking. (b) General cracking. Source: Ref 8 The corrosion pattern on the test specimen, particularly the number and distribution of cracks, can impair the precision of results obtained by either constant-strain or constant-load tests. When isolated stress-corrosion cracks propagate in a specimen stressed by either method, the average tensile stress on the net section increases rapidly until the notch fracture strength is reached and the specimen breaks (Fig. 5a). Less penetration is required for fracture of specimens under dead load; this indicates that specimen life is shorter with lowerstiffness stressing frames. When microcracks initiate close to one another, their individual stress concentrations interact and are relaxed. Consequently, there may not be a sufficient stress concentration in the true constantstrain test to propagate further SCC, and the specimen will not break (Fig. 5b). Under a constant load, however, the growth of many cracks continues, and the specimen ultimately breaks. With general cracking, crack propagation can be strongly influenced by frame stiffness. Therefore, SCC comparison of specimens tested at stress levels just above their thresholds is complicated by random variations in the cracking pattern, particularly when tested with a relatively stiff stressing system. Although constant-load stressing appears to be advantageous for testing materials with relatively high resistance to SCC, difficulties arise when small-diameter specimens are utilized to avoid the use of massive loads or lever systems. In some test environments, highly stressed specimens may fail from general or pitting corrosion and an attendant increase in the effective stress. Such non-SCC failures complicate interpretation of test results, unless failure by SCC is confirmed by examination using a light microscope and/or a scanning electron microscope. Such extraneous failures are less likely to occur with specimens loaded under constant strain. Therefore, small test specimens, which are generally preferred for laboratory screening tests and research studies, must be used with caution when estimates of serviceability are required. To determine serviceability, larger specimens should be used, and a stressing system should be selected that best duplicates the anticipated service conditions.

Bending versus Uniaxial Tension Historically, the most extensively used stressing systems have incorporated constant-deformation specimens stressed by bending. This method is versatile because of the variety of simple techniques that can be used to test most metal products in all types of corrosive environments. The state of stress in a bend specimen, however, is

much more complex than in a tension specimen. Theoretically, tensile stress is uniform throughout the cross section in the tension specimen, except at corners in rectangular sections, but the tensile stress in bend specimens varies through the specimen thickness. Tensile stress is at a maximum on the convex surface and decreases steeply to zero at the neutral axis. It then changes to a compressive stress, which reaches a maximum on the concave surface. Thus, only about 50% of the metal surface is under tension, and stress can vary from maximum to zero, depending on the stressing system. As SCC penetrates the metal, the stress gradient through the section thickness produces changes in stresses and strains that are different from those in a uniaxial tension specimen. This tendency yields significantly different SCC responses for the two types of stressing (Fig. 6).

Fig. 6 Comparison of the SCC response with bending versus direct tension stressing under constant load for Al-5.3Zn-3.7Mg-0.3Mn-0.1Cr T6 temper alloy sheet. Tested to failure in 3% NaCl plus 0.1% H2O2. Source: Ref 9 Bending stress specimens experience other sources of variability in stress that are not present with direct tension stressing. Variations occur in the principal longitudinal stress across the width of the specimen as well as with the presence of biaxial stresses, both of which are influenced by the design of the specimen. Therefore, just as in the case of constant-load stressing, optimal control of stress and more severe testing conditions are provided by uniaxial tension stressing. Statically loaded, smooth test specimens for SCC tests can be divided into three general categories: elasticstrain specimens, plastic-strain specimens, and residual-stress specimens. The commonly used specimen geometries for each of these categories are discussed in the sections that follow.

Elastic-Strain Specimens To control the surface tensile stress applied by deformation loading, strain is usually restricted to the elastic range for the test material. The magnitude of the applied stress can then be calculated from the measured strain and modulus of elasticity. In constant-load stressing, the load typically is measured directly, and the stress is calculated by using the appropriate formula for the specimen configuration and the method of loading. Load cells or calibrated springs may be useful for applying and monitoring possible changes in load during the test. The commonly used types of specimens for tests under elastic- range stress are described in this section. Bent-beam specimens can be used to test a variety of product forms. The bent-beam configuration is primarily used for sheet, plate, or flat extruded sections, which conveniently provide flat specimens of rectangular cross section, but it is also used for cast materials, rod, pipe, or machined specimens of circular cross section. This method is applicable to specimens of any metal that are stressed to levels less than the elastic limit of the material; therefore, the applied stress can be calculated or measured accurately (ASTM G 39, Ref 10; and ISO 7539-2, Ref 11). Stress calculations by this method are not applicable to plastically stressed specimens. Bent- beam specimens are usually tested under constant-strain conditions, but constant-load conditions can also be used. In either case, local changes in the curvature of the specimen when cracking occurs result in changes in stress and strain during crack propagation. The “test stress” is taken as the highest surface tensile stress existing at the start of the test, that is, before the initiation of SCC.

Several configurations of bent-beam specimens and stressing systems are illustrated in Fig. 7 and are described in detail in ASTM G 39 (Ref 10) and in ISO 7539-2 (Ref 11). When specimens are tested at elevated temperatures, the possibility of stress relaxation should be investigated.

Bolt loaded double-beam specimen dimensions for various plate thicknesses t a b L S mm in. mm in. mm in. mm in. mm in. 0.125 100 4.0 50 2.0 250 10.0 305 12.0 3.2 0.25 100 4.0 50 2.0 250 10.0 305 12.0 6.4 0.375 120 4.75 90 3.5 330 13.0 380 15.0 9.5 0.5 120 4.75 90 3.5 330 13.0 380 15.0 13 0.75 140 5.5 150 6.0 430 17.0 480 19.0 19 1.0 150 6.0 200 8.0 510 20.0 560 22.0 25 1.5 165 6.5 305 12.0 635 25.0 685 27.0 38 Fig. 7 Specimen and holder configurations for bent-beam stressing. (a) Two-point loaded specimen. (b) Three-point loaded specimen. (c) Four-point loaded specimen. (d) Welded double-beam specimen. (e) Bolt-loaded double-beam specimen. Formula for stressing specimen (e): Δd = 2fa/3Et(3L - 4a), where Δd is deflection (in inches), f is nominal stress (in pounds per square inch), and E is modulus of elasticity (in pounds per square inch). Source: Ref 12 Two-point loaded specimens can be used for materials that do not deform plastically when bent to (L - H)/H = 0.01. The specimens should be approximately 25 × 250 mm (1 × 10 in.) flat strips cut to appropriate lengths to produce the desired stress after bending, as shown in Fig. 7(a). The maximum stress occurs at the midlength of the specimen and decreases to zero at specimen ends. Three-point loaded specimens are flat strips that are typically 25 to 51 mm (1 to 2 in.) wide and 127 to 254 mm (5 to 10 in.) long. The thickness of a specimen is usually dictated by the mechanical properties of the material and the available product form. The specimen should be supported at the ends and bent by forcing a screw (equipped with a ball or knife-edge tip) against it at a point halfway between the end supports, as shown in Fig. 7(b). In a three-point loaded specimen, the maximum stress occurs at the midlength of the specimen and decreases linearly to zero at the outer supports. Two- and four-point loaded specimens are often preferred over the three-point loaded specimen, because crevice corrosion often occurs at the central support of the three-point loaded specimen. Because this corrosion site is very close to the point of highest tensile stress, it may cathodically protect the specimen and prevent possible crack formation, or it may cause hydrogen embrittlement. Furthermore, the pressure of the central support at the point of highest load introduces biaxial stresses at the area of contact and can introduce tensile stresses where compressive stresses are normally present.

Four-point loaded specimens are flat strips that are typically 25 to 51 mm (1 to 2 in.) wide and 127 to 254 mm (5 to 10 in.) long. The thickness of a specimen is usually dictated by the mechanical properties of the material and the available product form. The specimen is supported at the ends and is bent by forcing two inner supports against it, as shown in Fig. 7(c). The two inner supports are located symmetrically around the midpoint of the specimen. In a four-point loaded specimen, the maximum stress occurs between the contact points of the inner supports; the stress is uniform in this area. From the inner supports, the stress decreases linearly toward zero at the outer supports. The four-point loaded specimen is preferred over the three-point and two-point loaded specimens, because it provides a large area of uniform stress. Welded double-beam specimens consist of two flat strips 25 to 51 mm (1 to 2 in.) wide and 127 to 254 mm (5 to 10 in.) long. The strips are bent against each other over a centrally located spacer until both ends touch. The strips are held in position by welding the ends together, as shown in Fig. 7(d). In a welded double-beam specimen, the maximum stress occurs between the contact points of the spacer; the stress is uniform in this area. From contact with the spacer, the stress decreases linearly toward zero at the ends of the specimen, similar to a four-point loaded specimen. A bolt-loaded double-beam specimen is shown in Fig. 7(e), along with suggested specimen dimensions for various thicknesses of plate and the formula for stressing such specimens (Ref 12). The beam deflections required to develop the intended tensile stress are calculated with the formula and are then applied by bolting the ends of the beams together. The deflections are measured with a dial gage to within ±0.0127 mm (±0.0005 in.). Thus, the error in stress application—if the beams are of homogeneous material and if the cross sections are uniform—is within 2%. The precision of the deflection measurement is within 0.5%, and the error in determining the modulus of elasticity, E, is within 1%. Constant-moment beam specimens are designed such that a constant moment exists from one end to the other when the specimen is bent in the manner shown in Fig. 8 (Ref 13). This bending produces equal stress along the length of the specimen. The width-to-thickness ratio is less than 4 so that biaxial stresses are eliminated.

Fig. 8 Bent beam designed to produce pure bending. Source: Ref 13 This type of specimen offers the advantage of a relatively large area of material under a uniform stress. Such specimens can be used when the dimensions of the specimen are too small for other bent-beam specimens—for example, when specimens are taken in the short-transverse direction in plate (see Fig. 9c). The elastic stress σ in the convex surface is calculated by using: (Eq 1) where h is the distance between inner edges of the supports, y is the maximum deflection between inner edges of the supports, t is the thickness of the specimen, and E is the modulus of elasticity.

Fig. 9 Typical tuning-fork SCC test specimens. (a) Source: Ref 24. (b) Source: Ref 1. (c) Source: Ref 25 C-Ring Specimens. As discussed in ASTM G 38 (Ref 14), the C-ring is a versatile, economical specimen for quantitatively determining the susceptibility to SCC of all types of alloys in a wide variety of product forms. It is particularly well suited for testing tubing and for making short- transverse tests on various product forms, as shown in Fig. 10. The sizes of C-rings can be varied over a wide range, but rings with outside diameters less than about 16 mm ( in.) are not recommended because of increased difficulties in machining and decreased precision in stressing. The C-ring specimen is also covered by ISO 7539-5 (Ref 15).

Fig. 10 Sampling procedure for testing various products with C-rings. (a) Tube. (b) Rod and bar. (c) Plate Source: Ref 14 The C-ring is typically a constant-strain specimen with tensile stress produced on the exterior of the ring by the tightening of a bolt centered on the diameter of the ring. However, an almost constant load can be developed by placing a calibrated spring on the loading bolt. C-rings can also be stressed in the reverse direction by spreading the ring and creating a tensile stress on the inside surface. These methods of stressing are shown in Fig. 11.

Fig. 11 Methods of stressing C-rings. (a) Constant strain. (b) Constant load. (c) Constant strain. (d) Notched C-ring; a similar notch could be used on the side of (a), (b), or (c). Source: Ref 14 Circumferential stress is of principal interest in the C-ring specimen. This stress is not uniform (Ref 16), as discussed previously in the section “Elastic-Strain Specimens” in this article. The stress varies around the circumference of the C-ring from zero at each bolt hole to a maximum at the middle of the arc opposite the stressing bolt. In a notched C-ring, a triaxial stress state is present adjacent to the root of the notch (Ref 17). For

all notches, the circumferential stress at the root of the notch is greater than the nominal stress and can generally be expected to be in the plastic range. Generally, the C-ring can be stressed with high precision. The most accurate stressing procedure consists of attaching circumferential and transverse electrical strain gages to the surface stressed in tension, followed by tightening the bolt until the strain measurements indicate the desired circumferential stress. The amount of compression required on the C-ring to produce elastic straining and the degree of elastic strain can be predicted theoretically. Therefore, C-rings can be stressed by calculating the deflection required to develop a desired elastic stress (ASTM G 38) (Ref 14). In notched specimens, a nominal stress is estimated using a ring outside diameter measured at the root of the notch and by taking into consideration the stressconcentration factor, Kt, for the specific notch. O-ring specimens (Fig. 12) are used to develop a hoop stress in a particular part—for example, a cylindrical die forging in which a critical end-grain structure associated with the parting plane of the forging exists only at the surface of the forging. A relatively large surface area of metal is placed under a uniform tensile stress, and the O-ring stressing plug assembly simulates service conditions in structures containing interference-fit components. Stressed O- rings have also been used to evaluate protective treatments for the prevention of SCC (Ref 18).

Fig. 12 O-ring SCC test specimen (a) and stressing plug (b). The O-ring is stressed by pressing it onto the plug, as shown in (c). An O-ring is stressed by pressing it onto an oversized plug that is machined to a predetermined diameter to develop the desired stress at the outside surface of the ring. The nominal dimensions of this specimen can be varied to suit the part being tested, but certain characteristics should be observed to achieve adequate control of the stresses. The ring width should not be more than four times the wall thickness in order to ensure maximum uniformity of the hoop stress from the centerline to the edges of the ring. The tensile stress varies through the thickness of the ring and is highest at the inside surface. Interference required for stressing an O-ring can be calculated by using: (Eq 2) where I is the interference (on the diameter) between the O-ring and the plug, E is the modulus of elasticity, ID is the inside diameter, OD is the outside diameter, and F is the circumferential stress desired on the outside surface. Additional information regarding the design and stressing of O-ring specimens is given in Ref 19. Tension Specimens. Specimens used to determine tensile properties in air are well suited and easily adapted to SCC, as discussed in ASTM G 49 (Ref 8) and ISO 7539-4 (Ref 20). When uniaxially loaded in tension, the stress pattern is simple and uniform, and the magnitude of the applied stress can be accurately determined. Specimens can be quantitatively stressed by using equipment for application of either a constant load, a constant strain, or an increasing load or strain. This type of test is one of the most versatile methods of SCC testing because of the flexibility permitted in the type and size of the test specimen, the stressing procedures, and the range of stress level. It allows the simultaneous exposure of unstressed specimens (no applied load) with stressed specimens and subsequent tension testing to distinguish between the effects of true SCC and mechanical overload.

A wide range of test specimen sizes can be used, depending primarily on the dimensions of the product to be tested. Stress-corrosion test results can be significantly influenced by the cross section of the test specimen. Although large specimens may be more representative of most structures, they often cannot be prepared from the available product forms being evaluated. They also present more difficulties in stressing and handling in laboratory testing. Smaller cross-sectional specimens are widely used. They have a greater sensitivity to SCC initiation, usually yield test results rapidly, and permit greater convenience in testing. However, the smaller specimens are more difficult to machine, and test results are more likely to be influenced by extraneous stress concentrations resulting from nonaxial loading, corrosion pits, and so on. Therefore, use of specimens less than about 10 mm (0.4 in.) in gage length and 3 mm (0.12 in.) in diameter is not recommended, except when testing wire specimens. Tension specimens containing machined notches can be used to study SCC and hydrogen embrittlement. The presence of a notch induces a triaxial stress state at the root of the notch, in which the actual stress will be greater by a concentration factor that is dependent on the notch geometry. The advantages of such specimens include the localization of cracking to the notch region and acceleration of failure. However, unless directly related to practical service conditions, the results may not be relevant. Tension specimens can be subjected to a wide range of stress levels associated with either elastic or plastic strain. Because the stress system is intended to be essentially uniaxial (except in the case of notched specimens), great care must be exercised in the construction of stressing frames to prevent or minimize bending or torsional stresses. The simplest method of providing a constant load consists of a deadweight hung on one end of the specimen. This method is particularly useful for wire specimens. For specimens of larger cross section, however, lever systems such as those used in creep-testing machines are more practical. The primary advantage of any deadweight loading device is the constancy of the applied load. A constant-load system can be modified by the use of a calibrated spring, such as that shown in Fig. 13. The proving ring, as used in the calibration of tension testing machines, has also been adapted to SCC testing to provide a simple, compact, easily operated device for applying axial load (Fig. 14). The load is applied by tightening a nut on one of the bolts and is determined by carefully measuring the change in ring diameter.

Fig. 13 Spring-loaded fixture used to stress 3.2 mm (0.125 in.) thick sheet tensile specimens in direct tension. Source: Ref 12

Fig. 14 Ring-stressed tension specimen for field testing. Source: Ref 1

Constant-strain SCC tests are performed in low-compliance tension-testing machines. The specimen is loaded to the required stress level, and the moving beam is then locked in position. Other laboratory stressing frames have been used, generally for testing specimens of smaller cross section. Figure 15(a) shows an exploded view of such a stressing frame, and Fig. 15(b) illustrates a special loading device developed to ensure axial loading with minimal torsion and bending of the specimen.

Fig. 15 Equipment for constant-strain SCC testing. (a) Constant-strain SCC testing frame. Exploded view (left) showing the 3.2 mm (0.125 in.) diam tension specimen and various parts of the stressing frame. Final stressed assembly (right). Source: Ref 21. (b) Synchronous loading device used to stress specimens. The specimen is loaded to a prescribed strain value determined from a clip-on gage. The applied stress is given by the product of the strain and the material elastic modulus. A stressed assembly and one assembled finger-tight ready for stressing are shown.

For stressing frames that do not contain any mechanism for the measurement of load, the stress level can be determined from measurement of the strain. However, only when the intended stress is below the elastic limit of the test material is the average linear stress (σ) proportional to the average linear strain (ε), σ/ε = E, where E is the modulus of elasticity. When tests are conducted at elevated temperatures with constant-strain loaded specimens, consideration should be given to the possibility of stress relaxation. When stress relaxation or creep occurs in the test specimen, some of the elastic strain is converted to plastic strain and the nominal applied test stress is reduced. This effect is particularly important when the coefficients of thermal expansion are different for the specimen and stressing frame. Frequently, nonmetallic (plastic) insulators are used between the specimen and stressing frame to avoid galvanic action. If such plastic insulators are part of the stress-bearing system, creep (even at room temperature) can significantly alter the applied load on the specimen. Even though eccentricity in loading can be minimized to levels acceptable for tension-testing machines, tensile stress around the circumference of round specimens and at the corners of sheet-type specimens varies to some extent. Several factors may introduce bending moments on specimens, such as longitudinal curvature and misalignment of threads on threaded-end round specimens. These factors have a greater effect on specimens with smaller cross sections. Tests should be made on specimens with strain gages affixed to the specimen surface around the circumference of 90° or 120° intervals to verify strain and stress uniformity and to determine if machining practices and stressing jigs are of adequate tolerance and quality. When SCC occurs, it generally results in complete fracture of the specimen, which is easy to detect. However, when testing relatively ductile materials at stress levels close to the threshold of susceptibility, fracture may not occur during the period of exposure. The presence of SCC in such cases must be determined by mechanical tests or by metallographic examination, as discussed previously. To study trends in SCC susceptibility, such as in alloy development research, it is often necessary to detect small differences in susceptibility. For this purpose, it is advantageous to use replicate sets of specimens stressed at several levels, including zero applied stress. The sets are then removed for metallographic examination or tension tests after appropriate periods of exposure. Figure 16 illustrates the use of this procedure with samples of 7075 aluminum alloy that have been given different thermal treatments to decrease susceptibility to SCC. Analysis of these breaking stress data by extreme value statistics enables calculation of survival probabilities and the estimation of a threshold stress, without depending on failures during exposure. By using an elastic-plastic fracture mechanics model, an effective flaw size is calculated from the mean breaking stress, the strength, and the fracture toughness of the test material. The effective flaw size corresponds to the weakest link in the specimen at the time of the tension test, and it therefore represents the maximum penetration of the SCC. An advantage to using flaw depth to examine SCC performance is that the effects of specimen size and alloy strength and toughness can be normalized. In contrast, the specimen lifetime and breaking strength are biased by those mechanical (non-SCC) factors.

Fig. 16 Mean breaking stress versus exposure time for short-transverse 3.2 mm (0.125 in.) diam aluminum alloy 7075 tension specimens tested according to ASTM G 44 at various exposure stress levels. Each point represents an average of five specimens. Source: Ref 3

Mean trends in the 207 MPa (30 ksi) exposure data for the three temper variants of aluminum alloy 7075 examined in Fig. 16 are shown in Fig. 17. These results clearly illustrate that the thermal treatments used to reduce the SCC susceptibility of the 7075-T651 decreased the SCC penetration (Ref 22). The equivalent performance of the 7075-T7X1 3.2 and 5.7 mm (0.125 and 0.225 in.) diam specimens is evident. In contrast, Fig. 18 shows the specimen biases in SCC ratings obtained by traditional pass-fail methods (Ref 23).

Fig. 17 Effect of temper on SCC performance of aluminum alloy 7075 subjected to alternate immersion in 3.5% NaCl solution at a stress of 207 MPa (30 ksi). Mean flow depth was calculated from the average breaking strength of five specimens subjected to identical conditions. Source: Ref 22

Fig. 18 Influence of specimen configuration on SCC test performance (alternate immersion in 3.5% sodium chloride per ASTM G 44). Aluminum alloy 7075-T7X51 specimens stressed 310 MPa (45 ksi); each point represents 60 to 90 specimens. Source: Ref 23 Tuning-fork specimens are special-purpose specimens with numerous modifications (Fig. 9). In Europe, the metal is strained into the plastic range, and stresses and strains are usually not measured (Ref 24, 26). In the United States, however, these specimens have been used with measured strains in the elastic and plastic ranges. Specimens of the type shown in Fig. 9(b) are convenient when a small self-contained specimen is required that will afford some insight into the applied stresses. Such a specimen is particularly well suited for testing thin plate material in the longitudinal or long-transverse direction while keeping the original mill-finished surface intact.

Tuning-fork specimens are stressed by closing the specimen tines and restraining them in the closed position with a bolt placed at the tine ends. The amount of closure is determined from Eq 3, which was derived from the data obtained with strain gages placed at the base of the tines on calibration specimens (Ref 1): S = A Δxt

(Eq 3)

where S is the maximum tension stress in the outer fiber of either tine, A is the calibration constant, Δx is the total amount of closure at the tine ends, and t is the thickness of the tines. The stress on tuning forks with straight tines is greatest in a small area at the base of the tines. In tuning forks with tapered tines, the maximum stress extends uniformly along the tapered section. Tuning forks must be given the same consideration with regard to biaxial stresses as other flexurally loaded specimens. The miniature tuning fork shown in Fig. 9(c) was devised to conduct short-transverse tests on sections that are too thin for tensile specimens or C-rings to be obtained (Ref 25). As with other tuning-fork specimens, the relationship between strain on the grooved surface and the deflection at the ends of the legs can be determined through the use of strain gages.

Plastic-Strain Specimens Many accelerated SCC tests are performed with plastically deformed specimens, because these specimens are simple and economical to manufacture and use. These specimens are convenient for multiple replication tests of self- stressed (fixed-deflection) specimens in all environments. Because they usually contain large amounts of elastic and plastic strain, they provide one of the most severe tests available for smooth SCC test specimens. Generally, the stress conditions are not known precisely. However, the anticipated high level of stress can be obtained consistently only if the precautions described for each type of specimen are observed. Another consideration is that the cold work required to form the test specimen can change the metallurgical condition and the SCC behavior of certain alloys. Tests of this type are primarily used as screening tests to detect large differences between the SCC resistance of one alloy in several environments, one alloy in several metallurgical conditions in a given environment, and different alloys in the same environment. These tests are sometimes claimed to be too severe and therefore unsuitable for many applications, but the stress conditions are nevertheless representative of the high locked-in fabrication and assembly stresses frequently responsible for SCC in service. U-bend specimens are rectangular strips bent approximately 180° around a predetermined radius and maintained in this plastically (and elastically) deformed condition during the test. Standardized test methods for this type of specimen are described in ASTM G 30 (Ref 27) and ISO 7539-3 (Ref 28). Bends slightly less than or greater than 180° are also used, but the term U- bend is generally applied to test specimens that are bent beyond their elastic limits. Figure 19 illustrates typical U-bend configurations showing several different methods of maintaining the applied stress.

Alternativ e size A

L m m 80

in. 3.2

M m m 50

in. 2.0

W m m 20

in. 0.8

t m m 2.5

in. 0.09 8

D m m 10

in. 0.4

X m m 32

in. 1.2 6

Y m m 14

in. 0.5 5

R m m 5

in. 0.2

B

100 4.0

90

3.5

9

3.0

0.12

7

20

0.3 5 0.8

C

120 4.7

90

3.5

1.5

0.06

8

D

130 5.1

100 4.0

15

0.6

3.0

0.12

6

E

150 5.9

140 5.5

15

0.6

0.8

0.03

3

310 12. 2 510 20. G 1 Note: α = 1.57 rad

250 9.8

25

13

25

13. 0 6.5

0.51

460 18. 1

0.9 8 0.9 8

0.26

13

F

0.2 8 0.3 1 0.2 4 0.1 2 0.5 1 0.5 1

25

0.9 8 1.4

38

1.7 7 61 2.4 0 105 4.1 3 136 5.3 5

32

35 45

1.5 0 1.4

16

0.6

16

0.6

13

20

1.2 6 0.8

90

3.5

32

165 6.5

76

0.5 1 0.3 5 1.2 6 3.0

35

9

Fig. 19 Typical U-bend SCC specimens. (a) Various methods of stressing U-bends. (b) Typical U-bend specimen dimensions. Source: Ref 27 U-bend specimens can be used for all materials sufficiently ductile to be formed into a U- configuration without cracking. A U-bend specimen is most easily made from strips of sheet, but specimens can be machined from plate, bar, wire, castings, and weldments. Of primary interest in U-bend specimens is circumferential stress, which is not uniform, as discussed previously in the section on “Bent-Beam Specimens” in this article. Stress distribution in the U- bend specimen is discussed in detail in Ref 29. A good approximation of applied strain ε can be obtained by: (Eq 4) where t is the specimen thickness, and R is the radius of curvature at the point of interest. Knowledge of the stress-strain curve is necessary to determine the stress. When a U-bend specimen is formed, the material in the outer fibers of the bend is strained into the plastic portion of the true-stress/true-strain curve, such as in section AB in Fig. 20(a). Several other stress-strain relationships that can exist in the outer fibers of a stressed U-bend test specimen are shown in Fig. 20(b) through (e). The actual relationship obtained depends on the method of stressing used.

Fig. 20 True-stress/true-strain relationships for stressed U-bends. See text for discussion of (a) to (e). Source: Ref 27

Stressing is usually achieved by a one- or two- stage operation. Single-stage stressing is accomplished by bending the specimen into shape and maintaining it in that shape. The two types of stress conditions that can be obtained by single- stage stressing are defined by point X in Fig. 20(b) and (c). In Fig. 20(c), some elastic-strain relaxation has occurred by allowing the U-bend legs to spring back slightly at the end of the stressing sequence. Two-stage stressing involves forming the approximate U-shape and then allowing the elastic strain to relax completely before the second stage of applying the test stress. The applied test strain can be a percentage (from 0 to 100%) of the tensile elastic strain that occurred during preforming (Fig. 20d) or can involve additional plastic strain (Fig. 20e). The convex specimen surface is stressed in tension in the region 0NM (Fig. 20d), and the concave surface is in compression. In the region MP, the situation is reversed; that is, compression is on the convex surface, and tension is on the concave surface. The slope MN of the curve shown in Fig. 20(d) is steep. Therefore, it is often difficult to apply reproducibly a constant percentage of the total elastic prestrain, and the specimen surface may remain under compressive stress. Accordingly, the stress conditions in Fig. 20(b) and (e) are recommended because they result in a more severe test (that is, higher applied stress). Thus, the final applied strain prior to testing consists of plastic and elastic strain. To achieve the conditions illustrated in Fig. 20(b) and (e), springback of the U-bend legs after achieving the final plastic strain must be avoided. For materials with relatively low creep resistance, there will be some strain relaxation. It is important that the net residual stress be tensile. Compressively stressed surfaces are not normally prone to SCC. In fact, shot peening is a practical method for mitigating SCC by imparting a residual compressive stress on the metal surface before exposure to the service environment.

Residual-Stress Specimens Most industrial SCC problems are associated with residual tensile stresses developed in the metal during such processes as heat treatment, fabrication, and welding. Therefore, residual- stress specimens simulating anticipated service conditions are useful for assessing the SCC performance of some materials in particular structures and in specific environments. Plastic Deformation Specimens. Residual stresses resulting from such fabricating operations as forming, straightening, and swaging that involve localized plastic deformation at room temperature can exceed the elastic limit of the material. Examples of specimens of this type that have been used are shown in Fig. 21 and 22. Other specimen types used include panels with sheared edges, punched holes, or stamped identification numbers and specimens that show evidence of other practical fabricating operations.

Fig. 21 SCC test specimens containing residual stresses from plastic deformation. (a) Cracked cup specimen (Ericksen impression). Source: Ref 1. (b) Joggled extrusion containing SCC in the plastically deformed region. Source: Ref 9

Fig. 22 SCC test specimens containing residual stresses from plastic deformation. Shown are 12.7 mm (0.5 in.) diam stainless steel tubular specimens after SCC testing. (a) and (b) Annealed tubing that was cold formed before testing. (c) Cold-worked tubing tested in the as-received condition. Source: Ref 1 Weld Specimens. Residual stresses developed in and adjacent to welds are frequently a source of SCC in service. Longitudinal stresses in the vicinity of a single weld are unlikely to be as large as stresses developed in plastically deformed weldments, because stress in the weld metal is limited by the yield strength of the hot metal that shrinks as it cools. High stresses can be built up, however, when two or more weldments are joined into a more complex structure. Test specimens containing residual welding stresses are shown in Fig. 23. In fillet welds, residual tensile stress transverse to the weld can be critical, as indicated in Fig. 23(a) for a situation in which the tension stress acts in the short-transverse direction in an Al-Zn-Mg alloy plate.

Fig. 23 SCC test specimen containing residual stresses from welding. (a) Sandwich specimen simulating rigid structure. Note SCC in edges of center plate. Source: Ref 12. (b) Cracked ring-welded specimen. Source: Ref 1

References cited in this section 1. A.W. Loginow, Stress Corrosion Testing of Alloys, Mater. Prot., Vol 5 (No. 5), 1966, p 33–39

3. D.O. Sprowls et al., “A Study of Environmental Characterization of Conventional and Advanced Aluminum Alloys for Selection and Design: Phase II—The Breaking Load Test Method,” Contract NASI- 16424, NASA Contractor Report 172387, Aug 1984 6. R.N. Parkins, Stress Corrosion Test Methods—Physical Aspects, The Theory of Stress Corrosion Cracking in Alloys, J.C. Scully, Ed., NATO Scientific Affairs Division, 1971, p 449–468 7. G. Vogt, Comparative Survey of Type of Loading and Specimen Shape for Stress Corrosion Tests, Werkst. Korros., Vol 29, 1978, p 721–725 8. “Practice for Preparation and Use of Direct Tension Stress-Corrosion Test Specimens,” G 49, Metal Corrosion, Erosion, and Wear, Vol 03.02, Annual Book of ASTM Standards, American Society for Testing and Materials 9. H.L. Craig, Jr., D.O. Sprowls, and D.E. Piper, Stress-Corrosion Cracking, Handbook on Corrosion Testing and Evaluation, W.H. Ailor, Ed., John Wiley & Sons, 1971, p 259 10. “Practice for Preparation and Use of Bent- Beam Stress-Corrosion Test Specimens,” G 39, Metal Corrosion, Erosion, and Wear, Vol 03.02, Annual Book of ASTM Standards, American Society for Testing and Materials 11. “Corrosion of Metals and Alloys—Stress Corrosion Testing—Part 2: Preparation and Use of Bent-Beam Specimens,” ISO 7539-2, International Organization for Standardization 12. M.B. Shumaker et al., Evaluation of Various Techniques for Stress Corrosion Testing Welded Aluminum Alloys, Stress Corrosion Testing, STP 425, American Society for Testing and Materials, 1967, p 317–341 13. R.A. Davis, Stress Corrosion Cracking Investigation of Two Low Alloy, High Strength Steels, Corrosion, Vol 19 (No. 2), 1963, p 45t–55t 14. “Practice for Making and Using C-Ring Stress-Corrosion Test Specimens,” G 38 Metal Corrosion, Erosion, and Wear, Vol 03.02, Annual Book of ASTM Standards, American Society for Testing and Materials 15. “Corrosion of Metals and Alloys—Stress Corrosion Testing—Part 5: Preparation and Use of C-Ring Specimens,” ISO 7539- 5, International Organization for Standardization 16. S.O. Fernandez and G.F. Tisinai, Stress Analysis of Un-notched C-rings Used for Stress Cracking Studies, J. Eng. Ind., Feb 1968, p 147–152 17. F.S. Williams, W. Beck, and E.J. Jankowsky, A Notched Ring Specimen for Hydrogen Embrittlement Studies, Proc. ASTM, Vol 60, American Society for Testing and Materials, 1960, p 1192 18. D.O. Sprowls et al., “Investigation of the Stress Corrosion Cracking of High Strength Aluminum Alloys,” Final Technical Report for U.S. Government NASA Contract NAS-8-5340, Control No. 1-450-001167-01(lf), CPB-02-1215-64, 1967 19. Report of Task Group 1, of ASTM Subcommittee B-3/X, Stress Corrosion Testing Methods, Stress Corrosion Testing, STP 425, American Society for Testing and Materials, 1967, p 3–20; Proc. ASTM, Vol 65, 1965, p 182–197 20. “Corrosion of Metals and Alloys—Stress Corrosion Tests—Part 4: Preparation and Use of Uniaxially Loaded Tension Specimens,” ISO 7539-4, International Organization for Standardization

21. B.W. Lifka and D.O. Sprowls, Stress Corrosion Testing of 7079-T6 Aluminum Alloy in Various Environments, Stress Corrosion Testing, STP 425, American Society for Testing and Materials, 1967, p 342–362 22. R.J. Bucci et al., The Breaking Load Method: A New Approach for Assessing Resistance to Growth of Early Stage Stress Corrosion Cracks, Corrosion Cracking, V.S. Goel, Ed., Proc. Int. Conf. and Exposition on Fatigue, Corrosion Cracking, Fracture Mechanics, and Failure Analysis, American Society for Metals, 1986, p 267–277 23. D.O. Sprowls et al., Evaluation of a Proposed Standard Method of Testing for Susceptibility to SCC of High Strength 7XXX Series Aluminum Alloy Products, Stress- Corrosion—New Approaches, STP 610, H.L. Craig, Jr., Ed., American Society for Testing and Materials, 1976, p 3–31 24. “Testing of Light Metals, Stress Corrosion Test,” German Standard DIN 50908, 1964 25. F.H. Haynie et al., “A Fundamental Investigation of the Nature of Stress Corrosion Cracking in Aluminum Alloys,” Technical Report AFML 66-267, USAF Contract No. AF 33(615)-1710 Air Force Materials Laboratory, June 1966 26. P. Brenner, Realistic Stress Corrosion Testing, Metallurgy, Vol 23 (No. 9), 1969, p 879–886 27. “Practice for Making and Using U-Bend Stress-Corrosion Test Specimens,” G 30 Metal Corrosion, Erosion, and Wear, Vol 03.02, Annual Book of ASTM Standards, American Society for Testing and Materials 28. “Corrosion of Metals and Alloys—Stress Corrosion Testing—Part 3: Preparation and Use of U-Bend Specimens,” ISO 7539-3, International Organization for Standardization 29. H. Nathorst, Stress Corrosion Cracking in Stainless Steels, Part II—An Investigation of the Suitability of the U-Bend Specimen, Weld. Res. Counc. Bull., Series No. 6, Oct 1950


Static Loading of Precracked (Fracture Mechanics) Specimens The use of precracked (fracture mechanics) specimens is based on the concept that large structures with thick components are apt to contain cracklike defects. After a stress-corrosion crack begins to grow, or if the specimen is provided with a mechanical precrack, classical stress analysis is inadequate for determining the response of the material subjected to stress in the presence of a corrodent. The mechanical driving force for cracks can be measured with linear elastic fracture mechanics theory in terms of the crack-tip stress-intensity factor, K, which is expressed in terms of the remotely applied loads, crack depth, and test specimen geometry. At or above a certain level of K, SCC in a susceptible material will initiate and grow in certain environments, but below that level no measurable propagation is observed (Ref 30). The apparent threshold stress intensity for the propagation of SCC (assuming that crack nuclei form in a manner that cannot be described by fracture mechanics, such as localized corrosion) is designated KISCC (or

Kth). Therefore, in terms of linear elastic fracture mechanics theory, for a surface crack in a large plate remotely loaded in tension, the shallowest crack (of a shape that is long compared to its depth) that will propagate as a stress-corrosion crack is acr = 0.2(KISCC/ TYS)2, where TYS is tensile yield strength. Thus, a crack that is shallower than this critical value will not propagate under the given environmental conditions. The value of acr incorporates the SCC resistance, KISCC, and the contribution of stress levels (of the order of the yield strength) to SCC that are due to residual or assembly stresses in thick component sections (Ref 4). Therefore, the application of fracture mechanics does not provide independent information about SCC; it simply provides a usable method for treating the stress factor in the presence of a crack. When the rate of SCC propagation is determined and plotted as a function of Kt (the crack- opening mode), the test results for a highly susceptible alloy will exhibit the general trend shown in Fig. 3. Actual curves vary depending on the SCC resistance and fracture toughness of the alloy. Although precracking may shorten or modify the initiation period, it does not circumvent it. Therefore, this method of testing also requires arbitrary and sometimes long exposure periods.

Test Specimen Selection Almost all standard plane-strain fracture toughness test specimens can be adapted to SCC testing. These standard configurations should be used to ensure valid fracture analyses. Comprehensive discussions on SCC testing with precracked specimens can be found in Ref 31, 32, 33. Precracked specimens are shown in Fig. 24 where they are classified with respect to loading methods and the relationship with the stress-intensity factor as SCC propagates. Preparation of fracture toughness test specimens and test procedures are detailed in ASTM E 399 (Ref 34), E 1820 (Ref 35), and G 168 (Ref 36). Some exemplar test specimen dimensions and tolerances are given in Fig. 25(a)(a) to (c). Minor modifications to accommodate different loading arrangements and to facilitate mechanical precracking can be made to these configurations without invalidating the plane-strain constraints on the specimens. Figure 26 illustrates alternative chevron-notch and face-groove designs.

Fig. 24 Classification of precracked specimens for SCC testing. Asterisks denote commonly used configurations. Source: Ref 33 ASTM E 1681 (Ref 37) covers the determination of environment-assisted cracking threshold stress-intensity parameters, KIEAC and KEAC for metallic materials, using fatigue precracked beam or compact fracture specimens. Standard terminology relating to fracture testing is included in ASTM Standard E1823 (Ref 38). Cantilever bend specimens (Fig. 25(a)a), sometimes referred to as single-edge-notched cantilever bend specimens, have been used in constant-load tests (K-increasing) for characterizing high-strength steels and titanium alloys (Ref 39). Equations 5, 6, 7 are recommended (Ref 40, 41):

(Eq 5)

(Eq 6) (Eq 7) where e is the base of natural logarithm (2.718), x = [0.1426 + 11.92(a/W) - 17.42(a/W)2 + 15.84(a/W)3 2.235(a/W)4], y = [6.188 + 12.98(a/W) - 41.19(a/W)2 + 54.98(a/W)3 - 22.28(a/W)4]. M is the applied bending moment, B is the specimen thickness (face grooves, when present, may be accounted for by replacing B with , where Bn is the net thickness at the base of the face grooves; see Fig. 26b), W is the depth of the specimen, a is the depth of the notch plus crack, E is the modulus of elasticity, 2V0 is the total crack mouth opening displacement at the top face of the specimen, and VLL is the total crack mouth opening displacement measured at the point of load application, which will vary depending on the load arm length. Equation 5 is an expression for the stress intensity of a rectangular beam in pure bending and is valid over a wide range of a/W values. It applies to mode I loading only, however, and the usual tests include a mode II component from resulting shear stresses. Equations 6 and 7 were determined by fitting experimental compliance data for cantilever bend specimens with a polynomial equation expressing the natural log of the normalized compliance as a function of a/W. These experimental values are in excellent agreement with those determined from Eq 5 for pure bending, even though the stress state at the crack tip will differ for cantilever bending. It has been suggested that analyses using pure bending expressions related to compliance measurement are suitable for testing with the cantilever bend configuration (Ref 40). Crack growth measurements can be made with clip gage readings in conjunction with the crack- opening displacement calibrations given previously or by any other method that can be verified within ±0.127 mm (±0.005 in.). Examples of various methods are given in Ref 40 and 42. Modified compact specimens (K-decreasing or K-increasing), as shown in Fig. 25(b)(b), are frequently referred to as 1T-WOL (wedge-opening loaded) or modified WOL specimens. Although most frequently used with constant-displacement (bolt) loading (Ref 39, 43), these specimens have also been used with constant load (Ref 3, 44, 45). Equations 8, 9, 10, 11 can be used to calculate stress-intensity levels and normalized crack- opening displacements for fatigue precracking, for initiation of stress-corrosion testing, and for subsequent intervals during the test. These equations are based on boundary colocation values determined for this type of specimen configuration with face grooves and bolt loading (threaded bolt against a rigid loading tip) (Ref 40). The polynomial regression equation agrees with experimentally determined colocation values within 1% for 0.2 = a/W = 0.95:

(Eq 8)

(Eq 9) where x = [1.830 + 4.307(a/W) + 5.871(a0/ W)2 - 17.53(a0/W)3 + 14.57(a0/W)4]. (Eq 10) where y = [1.623 + 3.352(a0/W) + 8.205(a0/ W)2 - 19.59(a0/W)3 + 15.23(a0/W)4].

(Eq 11)

where z = [1.623 + 3.352(ai/W) + 8.205(ai/ W)2 - 19.59(ai/W)3 + 15.23(ai/W)4]. In Eq 8, Eq 9, Eq 10, Eq 11, KIo is the desired starting stress intensity, a0 is the starting crack length, P is the load calculated to develop KIo with measured a0, W is the net width of the specimen measured from the load line, KIi is the stress-intensity after time interval i, ai is the crack length after time interval i, and 2VLL is the total crack mouth opening displacement at the load line. All other quantities are as defined previously. Double-beam specimens (K-decreasing or K-increasing), which are also referred to as double-cantilever beam specimens, are similar to modified compact specimens, but because of their greater width or length, they are well suited for studying SCC growth rates over a greater range of KI values. The smaller height of these specimens (Fig. 25(c)c) allows more versatility in performing short-transverse tests from moderate thicknesses of material. Like compact specimens, double-beam specimens are generally used with constant-displacement (bolt) loading for convenience, but they can also be used with constant load. Bolt-loaded specimens used with a test procedure similar to that described in Ref 46 have been extensively employed for short-transverse tests of aluminum alloy products (Ref 45, 47, 48). Equation 12, Equation 13, Equation 14, Equation 15 are recommended for general use with double-beam specimens: (Eq 12)

(Eq 13)

(Eq 14)

(Eq 15) Equation 12 is an expression reported in Ref 49. The simplified Eq 15 provides more versatility with high accuracy for a wider range of specimen configurations and K values (crack growth) than equations previously published (Ref 46). Two early KI calibrations based on stress analysis (Ref 50) and compliance (Ref 51) are shown in Fig. 27 and are in excellent agreement. The shape of these curves can also be used as a design guide for preparing specimens. If the test must be completed in the shortest possible time, a0 should be short to capitalize on the fact that the rate of decrease of KI with crack extension is maximum for shallow cracks. However, if maximum accuracy is desired, a deeper crack (effective notch length M, Fig. 25(c)(c) should be chosen so that errors in crack length measurement do not cause significant errors in KI. In early work with aluminum alloys (Ref 46, 47) a relatively short effective notch length was used (a0/H ≈ 0.9), followed later by deeper notches (a0/H ≈ 1.2 to 2.2), all with a 2H value of 25.4 mm (1.0 in.) (Ref 3, 45, 47, 48). The recommended starting a/H value shown in Fig. 25(c)(c) is about 2 to 2.2, depending on the length of the precrack. Limited tests of a smaller beam height of 2H = 12.7 mm (0.5 in.) have shown little effect on the amount and rate of crack growth in aluminum alloy 7075 plate (Ref 52). An alternative double-cantilever beam specimen has been developed for testing relatively thin sections (typically 6.4 mm, or in., thick) of low-alloy steels (Ref 53). The specimen is stressed by forcing an appropriately dimensioned wedge into the slot. These specimens have been used to determine the effect of hardness of low-alloy steels on their resistance to SCC in environments containing hydrogen sulfide. Constant KI specimens are well suited for studying the mechanisms of SCC, because the stress intensity, KI, is not dependent on crack depth and can be neglected in kinetic studies. Other attractive features are the relatively simple expressions for stress intensity and compliance and the apparent retention of plane-strain conditions in thin plate and sheet specimens. The cost of specimen preparation and instrumentation, however, prohibits its use for extensive SCC characterizations. Reference 33 provides equations for the analysis of two types of constant KI specimens: the tapered doublebeam specimen and the doubletorsion-loaded single-edge cracked specimen. A recent evaluation of the double-torsion method (Ref 54) used Al-Zn-Mg alloy sheet 3.2 mm (0.125 in.) thick. By using the doubletorsion specimen, V-K curves were produced for aluminum alloy 7075-T651 sheet with conventional two-stage growth and plateau velocities that were only slightly higher than those for conventional double-cantilever beam tests of plate. Other precracked specimen configurations, such as those shown in Fig. 24, can be used for special testing conditions. Information on the preparation and use of these specimens and the related fracture mechanics equations are given in Ref 33 and 55, 56, 57.

Preparation of Precracked Specimens When using precracked SCC test specimens, the investigator must consider the dimensional (size) requirements of the specimen, its crack configuration and orientation, and machining and precracking of the specimen. These considerations are discussed in the sections that follow. Additional guidelines and recommendations on

specimen preparation in conjunction with fracture toughness testing are given in Ref 33, 34, 35, 36, 37 and 55, 56, 57. Dimensional Requirements. A basic requirement of all precracked specimen configurations is that the dimensions be sufficient to maintain predominantly triaxial stress (plane- strain) conditions, in which plastic deformation is limited to a very small region in the vicinity of the crack tip. Experience with fracture toughness testing has shown that for a valid KIc measurement neither the crack depth a nor the thickness B should be less than 2.5(KIc/YS)2, where YS is the yield strength of the material (Ref 34, 39). Because of the uncertainty regarding a minimum thickness for which an invariant value of KISCC can be obtained, guidelines for designing fracture mechanics test specimens should be tentatively followed for SCC test specimens. The threshold stressintensity value should be substituted for KIc in the aforementioned expression as a test of its validity. If specimens are to be used for determination of KISCC, the initial specimen size should be based on an estimate of the KISCC of the material. Overestimation of the KISCC value is recommended; therefore, a larger specimen should be used than may eventually be necessary. When determining stress-corrosion crack growth behavior as a function of stress intensity, specimen size should be based on the highest stress intensity at which crack growth rates are to be measured (substitute KIo in the 2.5(KIc/YS)2 expression). Notch Configuration and Orientation. For SCC testing, the depth of the initial crack-starter notch—that is, the machined slot with a fatigue or mechanical pop-in crack at its apex—can be as short as 0.2W. Guidelines for the depth of the notch depend on the limits of accurate KI calibration with respect to the range of a/W or a/H and the considerations discussed previously for double-beam specimens. Several designs of crack-starter notches are available for most plate specimens. The machined slot is used to simulate a crack, because it is impractical to produce plane cracks of sufficient size and accuracy in plate specimens. ASTM E 399 (Ref 34) recommends that the notch root radius should not be greater than 0.127 mm (0.005 in.), unless the chevron form is used, in which case it may be 0.25 mm (0.01 in.) or less (Fig. 26). This tolerance can be easily achieved with conventional milling and grinding equipment. A significant factor in the SCC testing of thick sections of some metals, such as aluminum and titanium, is the direction of applied stress relative to the grain structure. A standardized plan for identifying the loading direction, the fracture plane, and the direction of crack propagation is shown in Fig. 28. Machining. Specimens of the required orientation should be machined from products in the fully heat treated and stress-relieved condition to avoid complications due to residual stresses in the finished specimens. Safeguards against the presence of residual stresses are especially important for precracked specimens because these specimens are usually bulky and contain notches that are machined deep into the metal. For specimens of material that cannot easily be completely machined in the fully heat treated condition, the final thermal treatment can be given before the notching and finishing operations. However, fully machined specimens should be heat treated only when the heat treatment will not result in distortion, residual stress, quench cracking, or detrimental surface conditions. Precracking. Fatigue precracking should be done in accordance with ASTM E 399, E 1820, G 168, and E 1681 (Ref 34, 35, 36, 37, respectively). The K level used for precracking each specimen should not exceed about two-thirds of the intended starting K-value for the environmental exposure. This prevents fatigue damage or residual compressive stress at the crack tip, which may alter the SCC behavior, particularly when testing at a K level near the threshold stress intensity for the specimen. Aluminum alloy specimens can also be precracked by pop-in methods (wedge-opening loaded to the point of tensile overload), but steel and titanium alloys are usually too strong and tough to pop in without breaking off one of the specimen arms. Chevron notches are usually used to facilitate starting such mechanical precracks, and face grooves are sometimes necessary to produce straight precracks in tougher alloys (Fig. 26). These modifications may also be necessary to control fatigue precracking of some materials. When a specimen is mechanically precracked by pop in, the load should be maintained and should not be reduced for testing at a lower initial K-value. Reducing the load (crack mouth opening displacement) required for pop in will result in residual compressive stress at the crack tip, which could interfere with SCC initiation. When testing specimens at a relatively low fraction of KIc, fatigue precracking is recommended.

Testing Procedure For all methods using precracked specimens, the primary objective is usually to determine KISCC or Kth, threshold stress intensity for SCC for the alloy and environment combination. One procedure, similar to that

used with smooth specimens, depends on the initiation of SCC at various levels of applied KIo values. Both constant- load (K-increasing) and constant-displacement (K-decreasing) tests can be used. The latter procedure, which is unique to precracked specimens, involves crack arrest. This technique requires a K-decreasing constant-displacement test. These methods are compared in Fig. 29, which illustrates the shift in the stressintensity factor as SCC growth occurs. K-Increasing versus K-Decreasing Tests. In constant-load specimens (K-increasing tests), stress parameters can be quantified with confidence. Because crack growth results in an increasing crack opening, there is less likelihood that corrosion products will block the crack or wedge it open. Crack-length measurements can be made readily with several continuous-monitoring methods. A wide selection of constant-load specimen geometries are available to suit the test material, experimental facilities, and test objective. Therefore, crack growth can be studied under either bend or tension loading conditions. Specimens can be used to determine KISCC by the initiation of a stress-corrosion crack from a preexisting fatigue crack using a series of specimens or to measure crack growth rates. The principal disadvantages of constant-load specimens are the expense and bulk associated with the need for an external loading system. Bend specimens can be tested in relatively simple cantilever beam equipment, but specimens subjected to tension loading require constant- load creep-rupture equipment or similar testing machines. In this case, expense can be minimized by testing chains of specimens connected by loading links that are designed to prevent unloading upon failure of individual specimens. Because of the size of these loading systems, it is difficult to test constant-load specimens under operating conditions, but they can be tested in environments obtained from operating systems.

Fig. 25(a) Proportional dimensions and tolerances for cantilever bend test specimens. Width = W; thickness (B) = 0.5W; half loading span (L) = 2W; notch width (N) = 0.065W maximum if W >25 mm (>1.0 in.); N = 1.5 mm (0.06 in.) maximum if W = 25 mm (1.0 in.); effective notch length (M) = 0.25 to 0.45W; effective crack depth (a) = 0.45 to 0.55 W

Fig. 25(b) Proportional dimensions and tolerances for modified compact specimens. Surfaces should be perpendicular and parallel as applicable to within 0.002H TIR. The bolt centerline should be perpendicular to the specimen centerline within 1°. Bolt of material similar to specimen where practical; fine threaded, square or Allen head. Thickness = B; net width (W) = 2.55B; total width (C) = 3.20B; half height (H) = 1.24B; hole diameter (D) = 0.718B + 0.003B; effective notch length (M) = 0.77B; notch width (N) = 0.06B; thread diameter (T) = 0.625B

Fig. 25(c) Proportional dimensions and tolerances for double-beam specimens. “A” surfaces should be perpendicular and parallel as applicable to within 0.002H TIR. At each side, the point “B” should be equidistant from the top and bottom surfaces to within 0.001H. The bolt centerline (load line) should be perpendicular to the specimen centerline to within 1°. Bolt of material similar to specimen where practical; fine threaded, square or Allen head. Half height = H; thickness (B) = 2H; net width (W) = 10H minimum; total width (C) = W + T; thread diameter (T) = 0.75 H minimum; notch width (N) = 0.14H maximum; effective notch length (M) = 2H Constant-displacement specimens (K-decreasing tests) are self-loaded; therefore, external stressing equipment is not required. Their compact dimensions also facilitate exposure to operating service environments. They can be used to determine KISCC by the initiation of stress-corrosion cracks from the fatigue precrack, in which case a series of specimens must be used to bracket the threshold value. This can also be achieved by the arrest of a propagating crack, because under constant-displacement testing conditions stress intensity decreases progressively as crack propagation occurs. In this case, a single specimen suffices in principle; in practice, the use of several replicate specimens is recommended to assess variability in test results. Constant-displacement specimens are subject to several inherent disadvantages. Oxide formation or corrosion products can wedge the crack surfaces open, thus changing the applied displacement and load. Oxide formation or corrosion products can also block the crack mouth, thus preventing the entry of corrodent, and can impair the

accuracy of crack length measurements by electrical resistance methods. Applied loads can be measured only indirectly by displacement changes or by other sophisticated instrumentation. Crack arrest must be defined by an arbitrary crack growth rate below which it is impractical to measure cracks accurately (commonly about 1010 m/s, or 1.5 × 10-5 in./h). Loading Arrangements and Crack Measurement. To monitor crack propagation rate as a function of decreasing stress intensity when testing constant-displacement loaded specimens, two of the three testing variables must be measured—crack depth (ai) or load (Pi) and crack- opening displacement at the load line (VLL). Although crack initiation and growth can be detected from change in either load or crack length, load change is usually more sensitive to these conditions. Therefore, crack advance is easier to detect in specimens loaded in a testing machine, an elastic loading ring, or an instrumented bolt than in specimens loaded with a bolt or wedge. Figures 30(a)(a) and (b) illustrate typical loading arrangements for which load changes can be automatically monitored (Ref 3, 44, 58).

Fig. 26 Alternative chevron notch (a) and face grooves (b) for single-edge cracked specimens

Fig. 27 Configuration and KI calibration of a double-beam plate specimen. Normalized stress intensity KI plotted against a/H ratio. (W - a) indifferent, crackline-loaded, single-edge cracked specimen. Source: Ref 33

Fig. 28 Specimen orientation and fracture plane identification. L, length, longitudinal, principal direction of metal working (rolling, extrusion, axis of forging); T, width, long-transverse grain direction; S, thickness, short-transverse grain direction; C, chord of cylindrical cross section; R, radius of cylindrical cross section. First letter: normal to the fracture plane (loading direction); second letter: direction of crack propagation in fracture plane. Source: Ref 34

Fig. 29 Comparison of determination of KISCC by crack initiation versus crack arrest. (a) Constant-load test. (b) Constant crack-opening displacement test. a0 = depth of precrack associated with the initial stress intensity KIo; Vpl = plateau velocity Figure 31 illustrates an ultrasonic method of measuring crack length at the interior (midwidth and quarter widths) of a bolt-loaded double- beam specimen. This method provides a more accurate measure of crack depth

than visual measurements made on the specimen surfaces. Various other techniques have been used, such as measurement of beam deflection for cantilever beam specimens (Ref 40) and changes in electrical resistance. Such arrangements, however, require calibration. It is feasible and desirable to obtain crack length measurements with a precision of at least ±0.127 mm (±0.005 in.).

Fig. 30(a) Wedge-opening load specimen loaded with instrumented bolt. Source: Ref 58

Fig. 30(b) Ring-loaded wedge-opening load specimen test setup. Box to the left of loading rings contains analog signal conditioning for load and displacement signals. The digital data-acquisition system consists of a scanner connected to the analog load and displacement signals, a digital voltmeter, and a portable computer used to read and store data and to control the other instruments. Source: Ref 3

Fig. 31 Ultrasonic crack measurement system for double-beam specimens. Bolt-loaded specimen is mounted on translation stage at center. Ultrasonic transducer is located above specimen, and the oscilloscope at left indicates (left to right) the top of the specimen, the crack plane, and the bottom face reflection. Digital readouts of stage position and peak height for the crack front measurement used to make consistent positioning measurements are shown (right). This system has a crack growth resolution of approximately 0.127 mm (0.005 in.). Source: Ref 3 Exposure to Environment. When practical for laboratory accelerated testing, the test environment should be brought into contact with the specimen before it is stressed or immediately afterward; this enhances access of the corrodent to the crack tip to promote earlier initiation of SCC and to decrease variability in test results. Similarly, in certain cases, it may also be beneficial to introduce the corrodent even earlier, that is, during precracking. However, unless facilities are available to begin environmental exposure immediately after precracking, corrodent remaining at the crack tip may promote blunting due to corrosive attack. In addition, corrosion of the specimen surfaces in the small volume of the precrack or the advancing stress-corrosion crack will change the composition of the environment that is in contact with the crack tip and can significantly affect the test results. For example, hydrolysis reactions can drastically reduce the pH of the aqueous test environment (Ref 59) and can induce embrittlement of some steels by corrosion-product hydrogen. Selection of an appropriate test duration presents problems that vary with the testing system; this includes the alloy and metallurgical condition, the test environment, and the loading method. Errors in interpretation of the test results can be caused by test durations that are either too short or too long. The optimal length of exposure can be best approached through recognition of meaningful crack propagation rates. What is considered meaningful depends on the available precision of measurement of crack lengths and an acceptably low rate for the criterion of a stress-intensity threshold (Fig. 3). A problem also exists with the correlation of SCC crack growth rates in the laboratory test and in an anticipated service environment. The question leads ultimately to the intended application and a determination of what is a tolerable amount of SCC growth for a given length of time. Calculation of Crack Growth Rates. There are several procedures for calculating crack growth rate, da/dt, as a function of stress intensity from crack growth curves. The simplest is a graphical Δa/Δt technique that may incorporate smoothing of the a versus t curve (Ref 46, 47, 48). Another widely used approach is smoothing of the crack growth curve by computer techniques for curve fitting the entire a versus t curve by a multiple-term polynomial function (Ref 40). ASTM E 647 (Ref 60) is also a useful resource, although it is specifically for measuring fatigue crack growth rates. Other techniques include a secant method and an incremental polynomial method, in which derivatives of the smoothed crack growth curve are calculated at various points to determine instantaneous crack growth rates. Instantaneous growth rates are then plotted against the instantaneous stress intensities, KIi, at corresponding time intervals to obtain graphs similar to that shown in Fig. 3. A limited study of the aforementioned four methods of treating crack growth data is presented for a highstrength aluminum alloy in Ref 3. All of the methods used to calculate crack growth rates produced the same general results, which were difficult to interpret because of large amounts of scatter resulting from the use of

small crack growth increments. Moreover, the significance of such graphs is dubious when the corrosivity of the environment and the length of exposure can invalidate the estimate of K by causing gross corrosion-product wedging effects and/or crack branching. Reduction of crack length data becomes useless without prior subjective interpretation of crack length versus time curves. Allowances should be made for extraneous effects caused by erratic or apparent initiation of stresscorrosion crack growth, scatter in the measurement data due to excessive crack front curvature, multiple crack planes, crack-tip branching, and gross wedging caused by corrosion products. A simple method of comparing materials by using crack growth curves is based on average growth rates taken from an exposure time of zero to an arbitrary time that is sufficient to achieve significant crack extension in the most SCC-susceptible materials being compared (Ref 52). This method not only rapidly identifies materials with relatively low resistance to SCC, but also provides numerical test results for highly resistant materials that may not develop a KI versus da/dt curve with a definite plateau (see the section “Testing of Aluminum Alloys” in this article).

References cited in this section 3. D.O. Sprowls et al., “A Study of Environmental Characterization of Conventional and Advanced Aluminum Alloys for Selection and Design: Phase II—The Breaking Load Test Method,” Contract NASI- 16424, NASA Contractor Report 172387, Aug 1984 4. B.F. Brown, “Stress Corrosion Cracking Control Measures,” National Bureau of Standards Monograph 156, U.S. Department of Commerce, June 1977 30. R.P. Wei, S.R. Novak, and D.P. Williams, Some Important Considerations in the Development of Stress Corrosion Cracking Test Methods, Mater. Res. Stand., Vol 12, 1972, p 25 31. “Characterization of Environmentally Assisted Cracking for Design—State of the Art,” National Materials Advisory Board Report No. NMAB-386, National Academy of Sciences, 1982 32. B.F. Brown, The Application of Fracture Mechanics to Stress Corrosion Cracking, Met. Rev., Vol 13, 1968, p 171–183 33. H.R. Smith and D.E. Piper, Stress Corrosion Testing with Precracked Specimens, Stress Corrosion Cracking in High Strength Steels and in Titanium and Aluminum Alloys, B.F. Brown, Ed., Naval Research Laboratory, 1972, p 17–78 34. “Standard Method Test for Plane-Strain Fracture Toughness of Metallic Materials,” E 399, Annual Book of ASTM Standards, Vol 03.01, American Society for Testing and Materials 35. “Standard Test Method for Measurement of Fracture Toughness,” E 1820, Annual Book of ASTM Standards, Vol 03.01, American Society for Testing and Materials 36. “Standard Practice for Making and Using Precracked Double Beam Stress Corrosion Specimens,” G 168, Annual Book of ASTM Standards, Vol 03.02, American Society for Testing and Materials 37. “Standard Test Method for Determining Threshold Stress Intensity Factor for Environment-Assisted Cracking of Metallic Materials,” E 1681, Annual Book of ASTM Standards, American Society for Testing and Materials 38. “Standard Terminology Relating to Fatigue and Fracture Testing,” E 1823, Annual Book of ASTM Standards, American Society for Testing and Materials

39. J.A. Hauser, H.R.W. Judy, Jr., and T.W. Crooker, “Draft Standard Method of Test for Plane-Strain Stress-Corrosion-Cracking Resistance of Metallic Materials in Marine Environments,” NRL Memorandum Report 5295, Naval Research Laboratory, March 1984 40. W.B. Lisagor, Influence of Precracked Specimen Configuration and Starting Stress Intensity on the Stress Corrosion Cracking of 4340 Steel, Environment-Sensitive Fracture: Evaluation and Comparison of Test Methods, STP 821, S.W. Dean, E.N. Pugh, and G.M. Ugiansky, Ed., American Society for Testing and Materials, 1984, p 80–97 41. H. Tada, P. Paris, and G. Irwin, The Stress Analysis of Cracks Handbook, Del Research Corporation, 1973 42. J.A. Joyce, D.F. Hasson, and C.R. Crowe, Computer Data Acquisition Monitoring of the Stress Corrosion Cracking of Depleted Uranium Cantilever Beam Specimens, J. Test. Eval., Vol 8 (No. 6), 1980, p 293–300 43. S.R. Novak and S.T. Rolfe, Modified WOL Specimen for KISCC Environmental Testing, J. Met., Vol 4 (No. 3), 1969, p 701–728 44. J.G. Kaufman, J.W. Coursen, and D.O. Sprowls, An Automated Method for Evaluating Resistance to Stress-Corrosion Cracking with Ring-Loaded Precracked Specimens, Stress Corrosion—New Approaches, STP 610, H.L. Craig, Jr., Ed., American Society for Testing and Materials, 1976, p 94–107 45. C. Micheletti and M. Buratti, New Testing Methods for the Evaluation of the Stress- Corrosion Behavior of High-Strength Aluminum Alloys by the Use of Precracked Specimens, Symposium Proc., Aluminum Alloys in the Aircraft Industry (Turin, Italy), Oct 1976, Technicopy Ltd., 1978, p 149–159 46. M.V. Hyatt, Use of Precracked Specimens in Stress Corrosion Testing of High Strength Aluminum Alloys, Corrosion, Vol 26 (No. 11), 1970, p 487–503 47. D.O. Sprowls et al., “Evaluation of Stress Corrosion Cracking Susceptibility Using Fracture Mechanics Techniques,” Contract NAS 8-21487, Contractor Report NASA CR-124469, May 1973 48. R.C. Dorward and K.R. Hasse, “Flaw Growth of 7075, 7475, 7050, and 7049 Aluminum Plate in Stress Corrosion Environments,” Final Technical Report for U.S. Government Contract NAS 8-30890, Oct 1976; Corrosion, Vol 34 (No. 11), 1978, p 386–395 49. W.B. Fichter, The Stress Intensity Factor for the Double Cantilever Beam, Int. J. Fract., Vol 22, 1983, p 133–143 50. J.E. Srawley and B. Gross, Stress Intensity Factors for Crackline-Loaded Edge-Crack Specimens, Mater. Res. Stand., Vol 7 (No. 4), 1967, p 155–162 51. S. Mostovoy, R.B. Crosley, and E.J. Ripling, Use of Crackline Loaded Specimens for Measuring Plane Strain Fracture Toughness, J. Mater., Vol 2 (No. 3), 1967, p 661–681 52. D.O. Sprowls and J.D. Walsh, Evaluating Stress-Corrosion Crack Propagation Rates in High Strength Aluminum Alloys with Bolt Loaded Precracked Double Cantilever Beam Specimens, Stress Corrosion—New Approaches, STP 610, H.L. Craig, Jr., Ed., American Society for Testing and Materials, 1976, p 143–156 53. R.B. Heady, Evaluation of Sulfide Corrosion Cracking Resistance in Low Alloy Steels, Corrosion, Vol 33 (No. 3), 1977, p 98–107

54. T.L. Bond, R.A. Yeske, and E.N. Pugh, Studies of Stress Corrosion Crack Growth in Al-Zn-Mg Alloys by the Double Torsion Method, Environment-Sensitive Fracture: Evaluation and Comparison of Test Methods, STP 821, S.W. Dean, E.N. Pugh, and G.M. Ugiansky, Ed., American Society for Testing and Materials, 1984, p 128–149 55. J.C. Lewis and G. Sines, Ed., Fracture Mechanics: Fourteenth Symposium, Vol II, Testing and Application, STP 791, American Society for Testing and Materials, 1983 56. S.W. Dean, E.N. Pugh, and G.M. Ugiansky, Ed., Environment-Sensitive Fracture: Evaluation and Comparison of Test Methods, STP 821, American Society for Testing and Materials, 1984 57. J.H. Underwood, S.W. Freiman, and F.I. Baratta, Ed., Chevron-Notched Specimens: Testing and Stress Analysis, STP 855, American Society for Testing and Materials, 1984 58. W.B. Gilbreath and M.J. Adamson, Aqueous Stress Corrosion Cracking of High Toughness D6AC Steel, Stress-Corrosion—New Approaches, STP 610, H.L. Craig, Jr., Ed., American Society for Testing and Materials, 1976, p 176–187 59. B.F. Brown, Concept of Occluded Corrosion Cells, Corrosion, Vol 26 (No. 8), 1970, p 249 60. “Standard Test Method for Measurement of Fatigue Crack Growth Rates,” E 647, Annual Book of ASTM Standards, American Society for Testing and Materials


Dynamic Loading: Slow-Strain-Rate Testing An important method for accelerating the SCC process in laboratory testing involves relatively slow-strain-rate tension (SSRT) testing of a specimen during exposure to appropriate environmental conditions. In other words, the test specimen is stretched monotonically in axial tension at a slow rate until failure. This method is also known as constant extension-rate tensile (CERT) testing. The application of slow dynamic strain exceeding the elastic limit assists in the SCC initiation. This accelerated technique is consistent with the various proposed general mechanisms of SCC, most of which involve plastic microstrain and film rupture. Slow-strain-rate tests can be used to test a wide variety of product forms, including parts joined by welding. Tests can be conducted in tension, in bending, or with plain, notched, or precracked specimens. The principal advantage of slow-strain-rate testing is the rapidity with which the SCC susceptibility of a particular alloy and environment can be assessed. Slow-strain-rate testing is not terminated after an arbitrary period of time. Testing always ends in specimen fracture, and the mode of fracture is then compared with the criteria of SCC susceptibility for the test material. In addition to its timesaving benefits, generally less scatter occurs in the test results. Comprehensive discussions on the slow-strain-rate testing technique can be found in Ref 61, 62, 63, 64. Procedures for conducting SSRT tests are described in ASTM G 129 (Ref 65). Critical Strain Rate. The most significant variable in slow-strain-rate testing is the magnitude of strain rate. If the strain rate is too high, ductile fracture will occur before the necessary corrosion reactions can take place. Therefore, relatively low strain rates must be used. However, at too low a strain rate, corrosion may be prevented because of repassivation or film repair so that the necessary reactions of bare metal cannot be sustained, and SCC may not occur. Although typical critical strain rates range from 10 -5 to 10-7 s-1 depending on the alloy and environment system, the most severe strain rate must be determined in each case. The repassivation reaction that is observed at very low strain rates and that prevents the formation of anodic SCC does not occur when cracking is the result of embrittlement by corrosion-product hydrogen. This mechanistic difference can be used to distinguish between anodic SCC (active path corrosion) and cathodic SCC (hydrogen embrittlement) as shown in Fig. 32.

Fig. 32 Effect of strain rate on SCC and hydrogen-induced cracking. Source: Ref 66

The fastest strain rate that will promote SCC in a given system depends on crack velocity. Generally, the lower the SCC velocity, the slower the strain rate required. Applied strain rates known to have promoted SCC in metal/ environment systems are listed in Table 2. Table 2 Critical strain rate regimes promoting SCC in various metal/ environment systems System Applied strain rate, s-1 10-4 and 10-7 Aluminum alloys in chloride solutions 10-6 Copper alloys in ammoniacal and nitrite solutions Steels in carbonate, hydroxide, or nitrate solutions and liquefied ammonia 10-6 10-5 Magnesium alloys in chromate/chloride solutions 10-6 Stainless steels in chloride solutions 10-7 Stainless steels in high-temperature solutions 10-5 Titanium alloys in chloride solutions The most relevant strain rates for various aluminum alloys are shown in Fig. 33. These trends illustrate that slow-strain-rate tests should be performed in a strain-rate regime that is appropriate for the given alloy and environment system.

Fig. 33 Strain-rate regimes for studying SCC of various aluminum alloys. Corrodent: 3% sodium chloride plus 0.3% hydrogen peroxide. Source: Ref 64 Test Specimen Selection. Standard tension specimens (ASTM E 8) (Ref 67) are generally recommended for use with the specified conditions of gage lengths, radii, and so on, unless specialized studies are being conducted. For initially smooth specimens, the strain rate at the onset of the test is clearly defined: however, once cracks have initiated and grown, straining is likely to concentrate in the vicinity of the crack tip, and the effective strain rate is unknown. Rigorous solutions for determining the strain rate at crack tips or notches are not available, but effective strain rates are likely to be higher than for the same deflection rate applied to plain specimens. Notched or precracked specimens can be used to restrict cracking to a given location—for example, when testing the heat-affected zone associated with a weld. Notched or precracked specimens can also be used to restrict load requirements where bending, as opposed to tensile loading, may offer an added benefit. The section thickness or diameter of such specimens is usually relatively small, so the testing duration is short. Testing Equipment. Constant-strain-rate apparatus requirements include sufficient stiffness to resist significant deformation under the loads necessary to fracture the test specimens, a system to provide reproducible, constant strain rates over the range of 10-4 to 10-8 s-1, and a cell to contain the test solution. Auxiliary equipment is used to control environmental conditions and to record test data. The testing equipment can also be instrumented to record load-elongation curves, which is convenient when testing at various strain rates. A typical constantstrain-rate unit is shown in Fig. 34. Various types of corrosion cells may be required to control the test conditions for specific studies.

Fig. 34 Typical slow-strain-rate test apparatus. Source: Ref 63

In addition to uniaxial tensile units, cantilever constant-strain-rate apparatus has also been used in which an extension arm attached to a cantilever beam specimen is lowered at a constant rate. This technique has been successfully used to study SCC of low-carbon steel in carbonate- bicarbonate environments to determine crack velocity, critical strain rates, and inhibitor effectiveness (Ref 68). Additional information on slow-strain-rate testing equipment and procedures is available in Ref 56 and 61. Assessment of Results. Historically, the principal methods of SCC assessment derived from SSRT testing were based on time to failure, maximum gross section stress developed during the tension test, percent elongation, fracture energy (area bounded by the load-elongation curve), and reduction in area. Figure 35 depicts stresselongation curves that illustrate how stress-corrosion cracks influence the elongation to fracture as well as the maximum load.

Fig. 35 Nominal stress versus elongation curves for carbon-manganese steel in slow-strain-rate test in boiling 4 N sodium nitrate and in oil at the same temperature. Source: Ref 62 To eliminate non-SCC effects, parallel tests are conducted in an inert environment, and a ratio of the result obtained in the corrodent divided by the result obtained in the inert environment is commonly used as an index of SCC susceptibility. For example, in Fig. 33, higher SCC resistance is denoted by higher ductility ratios. Figure 36 shows a stress-corroded specimen containing many secondary stress-corrosion cracks and reduced ductility at fracture. Some alloys experience rapid deterioration of mechanical properties on contact with certain corrosive environments; any additional effect of applied straining can best be assessed by comparison with the behavior of unstrained specimens. Therefore, it is essential that the cause of environmental degradation be verified as SCC.

Fig. 36 Macrographs of two carbon steel specimens after slow-strain-rate tests conducted at a strain rate of 2.5 × 10-6 s-1 and 80 °C (180 °F). The ductility ratio in this example was 0.74 (original diameter: 2.54 mm, or 0.100 in.). Left: Ductile fracture in oil. Right: SCC in carbonate solution Slow-strain-rate testing is very efficient in comparing environments in terms of their capability to produce SCC, for example, in steels having similar metallurgical characteristics. However, such comparisons are difficult and not very reliable when applied to groups of steels with different characteristics (Ref 66). Slow-strain-rate testing as generally used does not provide data that can be used for design purposes. Recent work, however, has shown that average SCC velocities, threshold stresses, and threshold strain rates can be obtained with modified techniques combined with microscopy (Ref 62, 68, 69). For example, average SCC crack velocities can be determined from the depth of the largest crack measured on the fracture surfaces of specimens that have failed completely, or in longitudinal sections on the diameter of specimens that have not experienced total failure, divided by the time of testing. With this procedure, SCC is assumed to initiate at the start of the test, which is not always true. With precracked specimens, other methods can be used to monitor crack growth and thus allow determination of crack velocities. The SCC behavior of a pipeline steel (Fig. 37) has been studied by using a precracked cantilever bend specimen in terms of threshold strain rate for crack growth and also in terms of crack growth rates analogous to the stage II plateau velocity shown in Fig. 3. Material properties, such as strength and toughness, that influence SCC performance when measured by tension testing are eliminated as factors; therefore, valid comparisons can be made of alloys with widely different structures and mechanical properties. Additional information on this method of assessment and the effects of strain rate can be found in Ref 70, 71, 72. Screening of alloys for sour oilfield service by the SSRT technique is covered by a NACE International test method (Ref 73). Reference 74 describes limitations of SSRT testing for evaluating SCC in the chemicalprocess industry.

Fig. 37 Effects of beam deflection rate on stress-corrosion crack velocity in precracked cantilever bend specimens of a carbon-manganese steel. Tested in a carbonate-bicarbonate solution at 75 °C (165 °F) and at a potential of -650 mV versus SCE. Source: Ref 62

References cited in this section 56. S.W. Dean, E.N. Pugh, and G.M. Ugiansky, Ed., Environment-Sensitive Fracture: Evaluation and Comparison of Test Methods, STP 821, American Society for Testing and Materials, 1984 61. G.M. Ugiansky and J.H. Payer, Ed., Stress Corrosion Cracking—The Slow Strain- Rate Technique, STP 665, American Society for Testing and Materials, 1979 62. R.N. Parkins, Development of Strain-Rate Testing and Its Implications, Stress-Corrosion Cracking— The Slow Strain-Rate Technique, STP 665, G.M. Ugiansky and J.H. Payer, Ed., American Society for Testing and Materials, 1979, p 5–25 63. J.H. Payer, W.E. Berry, and W.K. Boyd, Constant Strain Rate Technique for Assessing Stress-Corrosion Susceptibility, Stress-Corrosion—New Approaches, STP 610, H.L. Craig, Jr., Ed., American Society for Testing and Materials, 1976, p 82–93 64. N.J.H. Holroyd and G.M. Scamans, Slow- Strain-Rate Stress Corrosion Testing of Aluminum Alloys, Environment-Sensitive Fracture: Evaluation and Comparison of Test Methods, STP 821, S.W. Dean, E.N. Pugh, and G.M. Ugiansky, Ed., American Society for Testing and Materials, 1984, p 202–241 65. “Practice for Slow Strain Rate Testing to Evaluate the Susceptibility of Metallic Materials to Environmentally Assisted Cracking,” G 129, Metal Corrosion, Erosion, and Wear, Vol 03.02, Annual Book of ASTM Standards, American Society for Testing and Materials 66. C.D. Kim and B.E. Wilde, A Review of Constant Strain-Rate Stress Corrosion Cracking Test, Stress Corrosion Cracking—The Slow Strain Rate Technique, STP 665, G.M. Ugianski and J.H. Payer, Ed., American Society for Testing and Materials, 1979, p 97–112 67. “Standard Methods of Tension Testing of Metallic Materials,” E 8, Annual Book of ASTM Standards, Vol 03.01, American Society for Testing and Materials

68. R.N. Parkins, “Fifth Symposium on Line Pipe Research,” Catalog No. L30174, U1- 40, American Gas Association 69. W.R. Wearmouth, G.P. Dean, and R.N. Parkins, Role of Stress in the Stress Corrosion Cracking of a Mg-Al Alloy, Corrosion, Vol 29 (No. 6), 1973, p 251–258 70. R.N. Parkins and Y. Suzuki, Environment Sensitive Cracking of a Nickel-Aluminum Bronze under Monotonic and Cyclic Loading Conditions, Corros. Sci., Vol 23 (No. 6), 1983, p 577–599 71. R.N. Parkins, A Critical Evaluation of Current Environment-Sensitive Fracture Test Methods, Environment-Sensitive Fracture: Evaluation and Comparison of Test Methods, STP 821, S.W. Dean, E.N. Pugh, and G.M. Ugiansky, Ed., American Society for Testing and Materials, 1984, p 5–31 72. J. Yu, N.J.H. Holroyd, and R.N. Parkins, Application of Slow-Strain-Rate Tests to Defining the Stress for Stress Corrosion Crack Initiation in 70/30 Brass, Environment-Sensitive Fracture: Evaluation and Comparison of Tests Methods, STP 821, S.W. Dean, E.N. Pugh, and G.M. Ugiansky, Ed., American Society for Testing and Materials, 1984, p 288–309 73. “Slow Strain Rate Test Method for Screening Corrosion-Resistant Alloys (CRAs) for Stress Corrosion Cracking in Sour Oilfield Service,” NACE TM-01-98, NACE International, 1998 74. J.A. Beavers and G.H. Koch, Limitations of the Slow Strain Rate Test for Stress Corrosion Cracking Testing, MTI Publication No. 39, Materials Technology Institute of the Chemical Process Industries, 1994


Selection of Test Environments The primary environmental factors in SCC testing are the nature and concentration of anions and cations in aqueous solutions, electrochemical potential, solution pH, the partial pressure and nature of species in gaseous mixtures, and temperature. Separately or in combination, environmental variables can have a profound effect on the thermodynamics and kinetics of the electrochemical processes that control environmentally assisted fracture. Therefore, the choice of environmental conditions provides an important basis for developing accelerated SCC test methods. The environmental requirements for SCC vary with different alloys. Although a mechanical precrack or a critical strain rate provides a worst case for SCC from a mechanical standpoint, there does not appear to be a generally applicable worst case from an environmental standpoint. However, because the presence of moisture and saltwater is universal, the SCC characteristics of alloys in these environments—as well as in any special environment a given engineering structure may experience—are always of interest. Figure 38 illustrates that electrochemical factors can override mechanical factors in determining SCC initiation sites. Three cantilever beam specimens of PH13-8Mo stainless steel were tested in saltwater. Specimen A was tested at a high K level. With the participation of the chloride ions, the protective oxide film ruptured at the bottom of the precrack and initiated SCC, which was halted before the beam fractured completely. Specimen B was loaded at a lower K level. After 1300 h, a stress-corrosion crack initiated, but not in the precrack. Crack

initiation occurred under the wall of the cell that surrounded the central portion of the specimen and contained the saltwater.

Fig. 38 Cantilever beam specimens of PH13-8Mo stainless steel after testing. Experiments demonstrate that electrochemical factors can override mechanical factors in determining initiation sites of SCC. See text for details. Source: Ref 75 Careful examination of this specimen and replicate specimens revealed small crevice-corrosion pits under the wall that initiated SCC in an almost smooth surface. Even if these small pits had been as sharp as a fatigue crack, the K level would have been much lower than at the machined and fatigued notch. In the stagnant situation under the cell wall, the stainless steel reacted with the saltwater to form hydrochloric acid and other corrosion products from the metal. Therefore, the low pH in a crevice, due to the hydrolysis of chromium corrosion product, overcame the mechanical disadvantage of the lack of a precrack. Specimen C was then tested to verify the effectiveness of electrochemical conditions in crack initiation. Saturated ferric chloride was selected to lower the pH to the range inside an active corrosion pit in the stainless steel; application of the solution to the unnotched beam resulted in the immediate initiation of many cracks in the smooth surface. Hydrochloric acid was found to be equally effective. Stress-corrosion tests can be divided into two broad environmental classes: those conducted in actual service environments and those conducted under laboratory conditions. Service Environments and Field Testing. The following examples illustrate the value, and in some cases the necessity, of exposure tests performed in actual service environments as an adjunct to laboratory evaluation. The standard 3.5% sodium chloride alternate immersion test data for aluminum alloys 2024 and 7075 proved useless in predicting the serviceability of these aluminum alloys for handling rocket propellant oxidizers such as nitrogen tetroxide and inhibited red fuming nitric acid (Ref 76). The alternate immersion test showed 2024T351 and 7075- T651 to be susceptible to SCC at low short-transverse stresses, but 2024-T851 and 7075-T7351 were quite resistant even at high stresses. These data were supported by outdoor field tests in seacoast and industrial atmospheres.

However, in proof tests consisting of exposure to the actual service environment—inhibited red fuming nitric acid at 74 °C (165 °F)—SCC occurred in both tempers of 7075 alloy and did not occur in either temper of 2024 alloy (Fig. 39). There were no unexpected failures with the 2219-T87 and 6061-T651 materials, however.

Fig. 39 SCC resistance of various aluminum alloys in inhibited red fuming nitric acid versus alternate immersion in 3.5% NaCl solution. Each bar graph represents an individual short-transverse C-ring test specimen machined from rolled plate and stressed at the indicated level. Source: Ref 76 Simulated-service tests should be conducted under conditions duplicating the service environment exactly, as illustrated by the following example with Ti-6Al-4V alloy pressure tanks for propellant-grade nitrogen tetroxide (25 to 30 wt% Cr) is required, generally, for good corrosion resistance to hot corrosion. Nickel-base alloys with both chromium and aluminum show further improvement in hot corrosion resistance. Acid-base oxide reactions with molten sulfate through the measurement of oxide solubilities as a function of Na2O activity in fused Na2SO4 were examined (Ref 56). Oxide solubility is dependent on Na2O activity, which also serves to rank the acid-base character of individual oxides. Other impurities, such as vanadium (≥0.4 ppm), phosphorus, lead, chlorides, and unburned carbon, can be involved in lowering salt melting temperatures, altering the sulfate activity, or changing the solution chemistry and acidity/basicity that leads to accelerating hot corrosion. Vanadic hot corrosion appears to be potentially more complex, because five compounds exist in the sodium, vanadium, oxygen system (Ref 57). The hightemperature reaction of sulfate and vanadium with ceramic oxides involved a Na 2O- V2O5 system that could be explained by Lewis acid-base chemistry (Ref 58). Basic zirconia (ZrO2)-stabilizing oxides, such as Y2O3, do not react with Na3VO4 (or 3Na2O-V2O5) but do react with the V2O5 component of NaVO3 (Na2O- V2O5) and V2O5 itself to form YVO4 (Ref 59). Acidic oxides, such as Ta2O5, react with the Na2O component of Na2VO4 and NaVO3 to form sodium tantalates and yield α-TaVO5 with V2O5. The vanadate that is most corrosive in the initiation of vanadic attack will depend on the acidity/basicity of the coating or alloy oxide. No reaction occurs when the acid-base properties of a stabilizing oxide are equal (Ref 60). The thermochemistry of vanadate and sulfate melts and reaction with different stabilizing oxides with SO3-NaVO3 was studied (Ref 61). High chromium content (>25 to 30% Cr) is required for good resistance to hot corrosion. Nickel alloys with both chromium and aluminum show improved hot corrosion resistance. However, inspection of phase stability diagrams for the systems M-Na-O-S, where M can be nickel, cobalt, iron, aluminum, or chromium, indicates that there are no combinations of melt basicity and oxygen activity where these metals, absent of a protective oxide film, are stable in contact with fused sodium sulfate (Ref 62). The relative hot corrosion resistance of a number of alloys has been evaluated in incinerator environments where hot corrosion can occur (Ref 38, 63, 64). Yttria-stabilized zirconia (YSZ) is attacked at high temperatures and destabilized by phosphorus impurities in fuel (Ref 65). The acid P2O5 reacts with basic Y2O3 to form the salt YPO4. Zirconia also synergistically reacts with sodium and P2O5 to form NaZr2(PO4)3. The YSZ thermal barrier coatings (TBCs) have been exposed to PbSO4-Na2SO4 molten salts without observable destabilization or reaction with this ceramic (Ref 66). However, lead, as PbO, appears to cause TBC failures by reacting with chromium in the NiCrAlY bond coat to form PbCrO4. Chloridation. Chlorides often accumulate rapidly on metallic surfaces of test samples. In one study, typical deposits contained 21 to 27% Cl when the flue gas contained 40 to 140 ppm HCl (Ref 67). Municipal wastes were characterized as having a 0.5% halide dry content, of which 60 wt% was derived from organic polymer sources (Ref 68). Chloride salts have melting temperatures as low as 175 °C (347 °F), which can act as fluxing agents that dissolve protective oxide films. High-temperature components exposed to air from a marine environment will be laden with chlorides. The following compounds may cause rapid corrosion of carbon steel, if present: Melting point Compound °C °F 246 474 Molten SnCl2 199 390 SnCl2 + NaCl 283 541 ZnCl2 347 Eutectic PbCl2/FeCl3 175 504 Eutectic ZnCl2/NaCl 262 Attack by halogens at elevated temperatures occurs through the volatility of the reaction products. Oxides that form in combustion gases will be porous and prone to fracture. Stainless steels are generally passive, but surface pitting may occur in chloride-containing environments. Nickel- base alloys can be expected to have superior corrosion resistance, as compared to stainless steel alloys (Ref 69). Clay containing aluminum silicate may inhibit chloride-related corrosion by raising the melting points of chloride salts through the formation of sodium aluminum silicates, which expel HCl and SO3 (Ref 70). Problems with chlorides can be mitigated if

plastics and other sources of halogens are removed or minimized from the waste stream. Increasing the oxygen content and adding water vapor has also reduced the corrosion rate of various alloys by chlorides in a simulated waste incinerator environment (Ref 67). Hydrogen chloride can be formed from the combustion from chlorinecontaining polymers and can dissociate into hydrogen and chlorine gases. Low pCl2 will generate a reducing environment. However, HCl can react with oxygen (generally, catalyzed by surface oxides) to generate water and possibly very high Cl2 pressures. The corrosion products of chloridation tend to be volatile and have low melting points. Other halogens may also affect metals and alloys in a manner similar to chlorides. Laboratory halogenization testing environments are divided into two groups: those possessing no measurable oxygen and those containing measurable oxygen. In the oxygen-free environments, liquid phases and volatile reaction products may lead to erratic corrosion attack. Halogen-plus-oxygen environments often exhibit paralinear behavior as oxide scale and volatile halide reactions occur simultaneously. Corrosion-resistance testing of materials is properly conducted in electrically heated furnaces using ceramic tubes with end caps containing pusher rods to manipulate the corrosion specimens under atmosphere. The end caps should be coated with a castable material to protect them from corrosion. Gas composition may be either purchased or synthesized from the components, using halogen-resistant electronic flow controllers or mixers. The specimens may be in the form of pins with dimensions similar to those mentioned for sulfidation testing. Gas flow streams of a total flow rate of 500 cm3/min (30.5 in.3/min) are considered adequate to ensure a consistent environment for all specimens. The effluent stream should be reacted with a caustic solution before venting to remove the reacted and unreacted halogens. At the completion of testing, the specimens should be examined metallographically for evidence of voiding or liquid-phase corrosion. Scale identification by x- ray diffraction or electron-dispersive spectrometry can be helpful in clarifying reaction mechanisms. Hydrogen Interactions. In selected high- temperature reactions, steam may decompose on metal surfaces to form hydrogen and oxygen. The resulting chemisorbed hydrogen may diffuse into the metal to an appreciable level. Loss in tensile ductility of steels and nickel-base alloys has been observed in gaseous environments with a total hydrogen content of 0.1 to 10 ppm at -100 to 700 °C (-150 to 1300 °F). Hydrogen attack occurs when hydrogen diffuses into the metal (typically steel) and reacts with the carbides to form methane, 4H + Fe3C → CH4 + 3Fe; the methane causes subsequent internal microcracks that lead to brittle rupture. The larger methane gas molecules also tend to concentrate at the grain boundaries. When methane gas pressures exceed the cohesive strength of the grains, a network of discontinuous, intergranular microcracks is produced. The reaction rate depends on the amount of carbon in the alloy, the hydrogen concentration, diffusion, total gaseous pressure, and temperatures in the range of 200 to 600 °C (400 to 1110 °F). Hydrogen damage has been observed in utility boilers at temperatures as low as 316 °C (600 °F) (Ref 71). Hydrogen damage can occur in high-strength alloys, resulting in loss of tensile ductility. Nickel-base alloys are much less susceptible to hydrogen damage than ferrous-base alloys. Nickel and nickel-base alloys are susceptible to attack in gaseous hydrogen environments. The same factors that affect hydrogen interactions in ferrous alloys are also operative for nickel-base alloys, although to a slightly lesser degree, because face-centered cubic metals have a greater number of slip planes and have lower solubilities for hydrogen than body-centered cubic metals (Ref 72). Hydrogen in nickel-base alloys may lead to intergranular, transgranular, or quasi- cleavage cracking. The Fe-Ni-Cr (Incoloy) and Inconel alloys show ductility reductions when exposed to hydrogen, particularly age-hardenable alloys (Ref 72). Molten Metals. Corrosion may cause dissolution of an alloy surface directly, by intergranular attack, or by leaching. Liquid metal attack may also initiate alloying, compound reduction, or interstitial or impurity reactions. Carbon and low-alloy steels are susceptible to various molten metals or alloys, such as brass, aluminum, bronze, copper, zinc, lead-tin solders, indium, and lithium, at temperatures from 260 to 815 °C (500 to 1500 °F). Plain carbon steels are not satisfactory for long-term use with molten aluminum. Stainless steels are generally attacked by molten aluminum, zinc, antimony, bismuth, cadmium, and tin (Ref 73). Nickel, nickel-chromium, and nickel-copper alloys generally have poor resistance to molten metals, such as lead, mercury, and cadmium. In general, nickel-chromium alloys also are not suitable for use in molten aluminum (Ref 74). There are no metals or alloys known to be totally immune to attack by liquid aluminum (Ref 75). Liquid metal embrittlement (LME) is a special case of brittle fracture that occurs in the absence of an inert environment and at low temperatures (Ref 76). Decreased stresses can reduce the possibility of failure in certain embrittling molten alloys. Stainless steels suffer from LME by molten zinc. Small amounts of lead will embrittle nickel alloys, but molybdenum additions appear to improve lead LME resistance. The selection of fabricating processes must be chosen carefully for nickel- base superalloys. Liquid metal embrittlement can

occur when brazing precipitation-strengthened alloys, such as Unified Numbering System (UNS) N07041 (Ref 77). Many nickel superalloys crack when subjected to tensile stresses in the presence of molten (B-Ag) brazing filler alloys. Molten salts are often involved in sulfidation, chloridation, hot corrosion, or high-temperature coatings, as discussed previously. The corrosiveness of the environment depends on the surface temperature and the condition of and/or the corrosive ingredients in the medium. Molten salts are employed in a number of applications, including metal heat treating; nuclear, fossil, and solar energy systems; reactive-metal extraction; high-temperature batteries; and fuel cells. Molten salts tend to flux the inherent scale that forms on heat-resistant alloys. In the presence of molten ash products, the oxide, even in oxidizing environments, becomes unstable and dissolves. Oxygen and water vapor tend to accelerate molten salt corrosion. Molten salt corrosion can take the macroscopic form of uniform thinning, pitting, or internal or intergranular attack. Alkali sulfates deposited on the fireside surfaces of boilers may react with SO3 or SO2 to form mixtures of alkali pyrosulfates (melting point: 400 to 480 °C, or 750 to 900 °F) or alkali- iron trisulfates (melting point: 550 °C, or 1020 °F) that cause fireside corrosion of reheater and superheater tubes (Ref 78). Molten sodium pyrosulfates (Na2S2O7) (melting point: 400 °C, or 750 °F) or potassium pyrosulfates (K2S2O7) (melting point: 10,000 h in oxidation or corrosion studies) (Ref 122). Discontinuous methods do present mass gain data per specimen and impose a thermal shock when the specimens are removed from measurement, but they provide a simple experiment to be exposed to complex environments and to evaluate a wide variety of alloys in a single test. Table 4 Recommended testing thermogravimetric analysis Environment

atmospheres

Gas composition

and

temperatures

for

testing

materials

by

Temperature °C °F 450–1200 850–2200 400–1000 750–1850 400–850 750–1550 300–600 570–1100 800–1100 1450–2000 800–1300 1450–2350 350–750 660–1380 400–700 750–1300 400–700 750–1300 400–700 750–1300

Air, 2.5% H2O Air Air, 2.5% H2O, 0.1–1% SO2 Flue gas Air, 2.5% H2O, 0.1–1% SO2, 0.05–0.1% HCl Waste incineration 0.1–1% H2S, balance H2 Sulfidizing environment 1% CH4, balance H2, dew-point -45 °C (-50 °F) Carburizing environment 90% N2, 10% H2, dewpoint -45 °C (-50 °F) Nitriding environment 25% CO, 73% H2, 2% H2O Metal dusting environment 0.1–1% H2S, 5% CO, 2.5% H2O, balance H2 Coal gasification (wet) 0.1–1% H2S, 70% CO, 2.5% H2O, 25% H2 Coal gasification (dry) Waste gasification (pyrolysis) 90% H2O, 5% H2, 5% CO, 0.1% HCl, 0.05% H2S Source: Ref 121 The ability of TGA to generate fundamental quantitative data from almost any class of materials has led to its widespread use in every field of science and technology. Key application areas are: • • • • •

Thermal stability: Related materials can be compared at elevated temperatures under the required atmosphere. The TG curve can help to elucidate decomposition mechanisms. Kinetic studies: A variety of methods exist for analyzing the kinetic features of all types of weight loss or gain, either with a view to predictive studies or to understanding the controlling chemistry. Materials characterization: The TG and DTG curves can be used to “fingerprint” materials for identification or quality control. Corrosion studies: Thermogravimetry provides an excellent means of studying oxidation or reaction with other reactive gases or vapors. Simulation of industrial processes: The thermobalance furnace may be thought of as a mini-reactor, with the ability to mimic the conditions in some types of industrial reactor.

•

Compositional analysis: By careful choice of temperature programming and gaseous environment, many complex materials or mixtures may be analyzed by selectively decomposing or removing their components. This approach is regularly used to analyze, for example, filler content in polymers, carbon black in oils, ash and carbon in coals, and the moisture content of many substances.

Differential thermal analysis (DTA) measures the temperature difference between a reactive sample and a nonreactive reference and is determined as a function of time, providing useful information about the temperatures, thermodynamics, and kinetics of reactions. The DTA is a dynamic method detecting the heat of reaction (exothermic or endothermic) of a physical or chemical change of the measured system. A sample to be analyzed and an inert substance are heated up in a furnace simultaneously under the same conditions. Within the sample and the inert substance there are mounted thermocouples that are connected against each other, so that the temperature difference between sample and inert substance is measured. Differential scanning calorimetry (DSC) measures temperature and heat flows associated with thermal transitions in a material. It is thus the most generally applicable of all thermal analysis methods, because every physical or chemical change involves a change in heat flow. Differential scanning calorimetry involves recording the energy necessary to establish a zero temperature difference between a substance and a reference material against either time or temperature as each specimen is subjected to an identical temperature regime in an environment heated or cooled at a controlled rate. The resulting DSC curve represents the amount of heat applied per unit time as the ordinate value against either time or temperature as the abscissa; this is related to the kinetics of the process (Ref 123). Temperature calibration is carried out by running standard materials, usually, very pure metals with accurately known melting points. Energy calibrations may be carried out by using either known heats of fusion for metals, commonly indium, or known heat capacities. Synthetic sapphire (corundum or aluminum oxide) is readily available as a heat capacity standard, and the values for this have been accurately determined over a wide temperature range. Typical purge gases are air and nitrogen, although helium is useful for efficient heat transfer and removal of volatiles. Argon is preferred as an inert purge when examining samples that can react with nitrogen. Experiments can also be carried out under vacuum or under high pressure, using instruments of the appropriate design. In the absence of any discrete physical or chemical transformations, the baseline signal in DSC is related to the heat capacity of the sample. Differential scanning calorimetry allows this parameter to be determined with good accuracy over a wide temperature range. The conventional approach is to compare the signal obtained for the sample above that given by an empty pan, with the signal obtained for a standard material, usually sapphire, under the same conditions. Careful experimental technique is required to obtain accurate results, but heat capacities can be routinely measured to an accuracy of better than ±1%. Other techniques are available for heat capacity measurement by DSC. Accurate data can be obtained in narrow temperature intervals by using a nonequilibrium pulse technique, which is particularly useful when measurements need to be made in a region constrained by adjacent complicating phase transitions. Less timeconsuming experiments than those described previously can generate data more rapidly but at the expense of accuracy and precision, which may be adequate for a given purpose. Modulated- temperature differential scanning calorimetry (MTDSC), a recent enhancement of DSC, routinely generates heat capacity data from a standard experiment and can, in fact, measure this in a nominally isothermal condition. However, the classical method is still recommended for the best-quality results. Use of DSC provides a cost-effective method for obtaining kinetic and thermodynamic data for condensed phase phenomena. Observable processes include simple phase transitions, characterization of polymorphism, and the kinetics and thermochemistry for a variety of complex reactions. Thermodynamic data for pure substances include melting and boiling points, heat capacity, heat of fusion, heat of solution, and heat of vaporization. The DSC-DTA curve may show an exothermic or endothermic peak. The enthalpy changes associated with the events occurring are given by the area under the peaks. In general, the heat capacity will also change over the region, and problems may arise in the correct assignment of the baseline. In many cases, the change is small, and techniques have been developed for reproducible measurements in specific systems. Some possible processes giving enthalpy peaks are listed in Table 5.

Table 5 Type of peaks related to different processes detected by differential thermal analysis and differential scanning calorimetry Process Solid-solid transition Crystallization Melting Vaporization Sublimation Adsorption Desorption Desolvation (drying) Decomposition Solid-solid reaction Solid-liquid reaction Solid-gas reaction Curing Polymerization Catalytic reactions

Exothermic X X … … … X … … X X X X X X X

Endothermic X … X X X … X X X X X X … … …

Burner Rig Testing Experiences in the 1960s and 1970s led to the discoveries of severe high-temperature corrosion in shipboard gas turbine engines that usually did not occur in aircraft turbine engines. Early observations noted severe corrosion attack on the first-stage blade and vane components of a shipboard marine gas turbine engine was sufficiently rapid to cause engine failure in several hundred hours (Ref 124). The ingestion of sea salt and the combustion of fuels containing some measure of sulfur by gas turbine engines operating in marine environments can lead to corrosion of hot section components, particularly turbine vanes (nozzles) and blades (buckets). This attack was documented in the early open literature as hot corrosion, as discussed earlier (Ref 124, 125, 126). Accelerated high-temperature corrosion was later observed in shipboard waste incinerators (Ref 41, 63). The laboratory technique that was found to most nearly approximate the operating conditions of a gas turbine engine was the burner rig (Ref 117). Burner rig tests provide a compromise between simple laboratory and field tests (Ref 96). Burner rigs are versatile: they may be used with very-low-sulfur fuel to simulate oxidation conditions at high temperatures; they may be used with 1 to 6 wt% S, vanadium, and/or highcarbon-residue fuels to simulate sulfidation and hot corrosion conditions; and they may be used with injection of NaCl to simulate hot corrosion conditions in a marine environment. Laboratory evaluations are employed as a sorting test for materials (alloys and coatings) resistant to hot corrosion, sulfidation, and other high-temperature gaseous corrosion reactions. The basic definition of a burner rig is a device that burns fuel to produce the hightemperature corrosive environment. There are several variations of burner rigs, each compromising certain aspects of a gas turbine environment while minimizing costs for testing (Ref 117). There are basically three types of burner rigs: low-velocity atmospheric pressure rigs, high-velocity atmospheric pressure rigs, and highvelocity, high-pressure rigs. Table 6 compares the test parameters of low-velocity atmospheric pressure, highvelocity atmospheric pressure, and high-velocity, high-pressure burner rigs in use (Ref 117).

Table 6 Typical operating parameters for burner rigs (low-velocity atmospheric pressure; high-velocity atmospheric pressure; high velocity, high pressure) simulating gas turbine environments Users

Type

Sea salt concentration, ppm

Low-velocity atmospheric pressure 10 DTNSRDC, INCO Ducted and General Electric Co.

Fuel

Gas velocity, m/s

Thermal cycle

Test time, h

Specimens Type

Number/test Exposure

1

5

1/24 h

300– 1000

Fin

42

Rotating carousel

30/1

1

…

1/1 h

500

Rod

…

Rotating carousel

Varies

28/1

1

0.1

1/22 h

300– 1000

mm 24 in.)

Rotating carousel

30/1

1

8

600

20

30/1

1

Rotating carousel Rotating carousel

1650– 2010

…

1

100 at nozzle, much lower at specimen …

Yes— variable 1/20 h

6.35 (0.25 pin Fin

900– 925 700 and 900

1650– 1700 1290 and 1650

0.3

900– 1100

…

850

1560

…

1

10

Air/fuel Pressure, ratio atm

Fuel sulphur content, wt%

Metal temperature °C °F

Marine diesel or JP-5 doped with tertiary butyl disulfide Natural gas (SO4 added -1.56 l h-1) BS 2869

0.4–2

700 and 900

1290 and 1650

30/1

…

…

…

1.0

Varies

JP-5

…

Johnson Matthey & Co. Ltd. (U.K.)

Ducted

24

National Physical Laboratory (U.K.)

Ducted

Naval Air Propulsion Center Pratt & Whitney Aircraft Group, United Technologies Corp. Rolls Royce Leavesdon (U.K.)

Ducted

20 continuous, 9000 intermittent 10

Ducted

20

Jet A (SO, 1.3 gas added with primary air)

Ducted

Measured as salt flux

Societé Nationale d'Etude et de Construction de Moteurs d'Aviation (SNECMA) (France)

…

1.9 (Na)

Aviationgrade kerosene …

100– 300

Aerofoil

7

None

100– 500

Pin blade

…

Rotating carousel

…

…

…

…

…

Users

Type

Sea salt concentration, ppm

High-velocity atmospheric pressure Ducted 0.1 Admirality Research Establishment (U.K.) Ducted 6 Avco/Lycoming Division 15 (Na) Brown Bovari and … CIE (Germany) 0.06 Cranfield Institute Ducted of Technology (U.K.) Ducted … DTNSROC

Howmet Turbine Components Corp.

Unducted 5 or 50

NASA, Lewis Research Center

Unducted 0.5 Na as NaCl

National Aerospace Lab (The Netherlands)

Unducted 0–20

Rolls Royce, Derby (U.K.)

Unducted 0.5–4

Solar Turbine

Ducted

3 and 35

Gas velocity, m/s

Thermal cycle

Test time, h

Specimens

0.76

165

1/25 h

2.5– 1200

6.35 (0.25 rod

1

215 90

Yes— complex …

1

65/1

1

120–130

12/100 h

100

1290 and 1650 1445– 1750

50/1

1

65–265

1/24 h

36/1 to 44/1

1

215

Yes— complex

700– 1260

1290– 2300

20/1 to 50/1

1

100 and 330

Cylinder or 1–8 wedge

0.003 0.003 … 10 K1A/low pressure (a) Iron tolerance equals manganese content of alloy times 0.032. (b) Iron tolerance equals manganese content of alloy times 0.021. (c) Magnesium content of AZ63 reported as 0.2% It should also be noted that the nickel tolerance depends strongly on the cast form, which influences grain size, with the low-pressure cast alloys showing just a 10 ppm tolerance for nickel in the as-cast (F) temper. Therefore, alloys intended for low-pressure cast applications should be of the lowest possible nickel level (Ref 8). The low tolerance limits for the contaminants in AM60 alloy when compared to AZ91 alloy can be related to the absence of zinc. Zinc is thought to improve the tolerance of magnesium-aluminum alloys for all three contaminants, but it is limited to 1 to 3% Zn because of its detrimental effects on microshrinkage porosity and its accelerating effect on corrosion above 3%. For magnesium-rare earth, -thorium, and -zinc alloys containing zirconium, the normal saltwater corrosion resistance is only moderately reduced when compared to high-purity magnesium and magnesium-aluminum alloys—0.5 to 0.76 mm/yr (20 to 30 mils/yr) as opposed to less than 0.25 mm/yr (10 mils/yr) in 5% salt spray— but contaminants again must be controlled. The zirconium alloying element is effective in this case because it serves as a strong grain refiner for magnesium alloys, and it precipitates the iron contaminant from the alloys before casting. However, if alloys containing more than 0.5 to 0.7% Ag or more than 2.7 to 3% Zn are used, a sacrifice in corrosion resistance should be expected (Fig. 4). Nevertheless, when properly finished these alloys provide excellent service in harsh environments. Heat Treating, Grain Size, and Cold-Work Effects. Heating influences the salt-spray corrosion rate of die-cast commercial magnesium-aluminum alloys as shown in Fig. 6, which shows that alloys with higher residualelement (iron, nickel, and copper) concentrations were more negatively impacted by temperature. Using controlled-purity AZ91 alloy cast in both high- and low-pressure forms, the contaminant-tolerance limits have been defined as summarized in Table 7 for the as-cast (F), the solution treated (T4, held 16 h at 410 °C, or 775 °F, and quenched), and the solution treated and aged (T6, held 16 h at 410 °C, or 775 °F, quenched, and aged 4 h at 215 °C, or 420 °F).

Fig. 6 Effect of heating temperature on corrosion rate of die-cast AZ91D and AM60B in salt-spray test for 10 days using ASTM B 117 method. Data are for test specimens that were heated from 0.5 to 36 h. Source: Ref 12 Table 7 Contaminant tolerance limits versus temper and cast form for AZ91 alloy High-pressure die cast, 5–10 μm average grain size; low- pressure cast, 100–200 μm average grain size Contaminant, % Critical contaminant limit(a) High pressure Low pressure F F T4 T6 0.032 Mn 0.032 Mn 0.035 Mn 0.046 Mn Iron 0.0050 0.0010 0.001 0.001 Nickel 0.040 0.040 99.5 0.05 Pb, 1.0 Zn, Cu + named elements > 99.5. If welded, see spec for additional limits. …

…

0.1– 0.2

1.8– 2.4

0.15– 0.25

0.15

0.5

…

N06030 G-30

43

2

15

…

…

0.03

…

1

…

0.3

21.0– 25.0

…

1.5– 4.0 …

…

45.0 min

4.0– 6.0 …

0.8

N06045 45TM

28.0– 31.5 26.0– 29.0

0.02 P, 0.02 S, 0.35 V 0.020 P, 0.010 S, 0.05–0.12 Y, 0.01–0.10 Zr …

…

…

…

0.05– 0.12

1.00

2.5– 3.0

…

N02201 201 N04400 400 N04405 R-405 N05500 K-500

…

0.1

…

0.05–0.12 N, 0.020 P, 0.010 S, total rare earth (RE) 0.05–0.15; approx 50% of RE is cerium

UNS Common No. name N06059 59 N06200 2000 N06600 600 N06601 601 N06617 617 N06625 625 N06686 686 N06690 690 N06985 G-3 N07214 214

N07725 725

Composition, % Ni Cr bal 22.0– 24.0 bal 22.0– 24.0 bal 14.00– 17.00 58.0– 21.0– 63.0 25.0 44.5 20.0– min 24.0 58.0 20.0– min 23.0 bal 19.0– 23.0 58.0 27.0– min 31.0 bal 21.0– 23.5 bal 15.0– 17.0

N07750 X-750

55.0– 59.0 bal

N07754 MA 754

78

N08020 20Cb-3

32.0– 38.0 35.0– 40.0 30.0– 34.0 30.0– 32.0

N08024 20Mo-4 N08028 28 N08031 3127hMo

19.0– 22.5 14.0– 17.0 19.0– 23.0 19.0– 21.0 22.5– 25.0 26.0– 28.0 26.0– 28.0

Cu …

Fe 1.5

Co 0.3

W …

Nb …

Ti …

…

…

…

…

Mo 15.0– 16.5 15.0– 17.0 …

1.3– 1.9 0.5 max 1

3.0

2.0

8

…

…

bal

…

…

…

…

0.3 max …

0.5

3

…

0.6

5

…

5.0

…

3.15– 4.15 …

0.4

… 0.5

7.0– 11.0 18.0– 21.0 2.0– 4.0

…

8.0– 10.0 8.0– 10.0 15.0– 17.0 …

…

…

10.0– 15.0 1

1.5– 2.5 …

5.0 2.0

3.0– 4.4 …

Al 0.1– 0.4 0.50

C 0.010

Mn 0.5

Si 0.10

B …

Other 0.015 P, 0.010 S

0.010

0.50

0.08

…

0.025 P, 0.010 S

…

0.08

…

0.5

…

…

1.0– 1.7 0.8– 1.5 0.4

0.1

1

0.5

…

…

0.05– 0.15 0.1

1

1

0.006 …

0.5

0.5

…

…

…

0.010

0.75

0.08

…

0.04 P, 0.02 S

…

0.02– 0.25 …

…

0.05

0.05

0.5

…

…

…

…

…

0.015

1.0

1.0

…

…

0.5

4.0– 5.0

0.05

0.5

0.2

1.00– 1.70 2.5

0.35

0.03

0.35

0.20

0.04 P, 0.03 S, Nb + Ta ≤ 0.50 0.006 0.05 Zr, 0.002– 0.040 Y, 0.015 P, 0.015 S … 0.015 P, 0.010 S

…

…

…

…

…

…

6.0– 8.0 0.5

1.5 max 0.5 max … …

2.75– 4.00 0.9

…

bal

…

…

5.0– 9.0 1

…

7.00– 9.50 …

…

…

…

…

0.5

0.3

0.05

…

…

…

bal

…

…

1

…

…

0.07

2

1

…

bal

…

…

…

0.03

1

0.5

…

…

…

…

0.15– 0.35 …

…

bal

…

…

0.03

2.50

1.00

…

bal

…

2.0– 3.0 3.5– 5.0 3.0– 4.0 6.00– 7.00

0.6 oxide …

…

…

…

…

0.015

2.0

0.05

…

0.030 P max, 0.030 S 0.03 P max, 0.005 S max,

… 3.0– 4.0 0.5– 1.5 0.6– 1.4 1.0– 1.4

yttrium

UNS No.

Common name

Composition, % Ni Cr Cu

N08330 RA-330

34.0– 37.0

N08800 800

30.0– 35.0 38.0– 46.0 23.0– 28.0 bal

17.0– 20.0

1.00

Fe

Co

Mo

W

Nb

Ti

Al

C

Mn

Si

B

bal

…

…

…

…

…

…

0.08

2.00

0.75– 1.50

…

19.0– … 39.5 … … … … 0.15– 0.15– 0.1 1.5 1 … 23.0 min 0.60 0.60 19.5– 1.5– bal … 2.5– … … 0.6– 0.2 0.05 1 0.5 … N08825 825 23.5 3.0 3.5 1.2 19.0– 1.00– bal … 4.00– … … … … 0.020 2.00 1.00 … N08904 904L 23.0 2.00 5.00 14.5– … 5.5 … 15.0– 3.0– … … … 0.01 … 0.08 … N10276 C-276 16.5 17.0 4.5 bal 1.0 … 2.0 … 26.0– … … … … 0.01 … 0.1 … N10665 B-2 max max 30.0 Composition is for identification only. Single values are maximum values, unless otherwise stated. UNS, Unified Numbering System

Other 0.15–0.25 N 0.03 P max, 0.03 S max, 0.025 Sn, 0.005 Pb … 0.04 P, 0.03 S 0.045 P max, 0.035 S … …

A.J. Sedriks, Corrosion Resistance of Stainless Steels and Nickel Alloys, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 697–702 Corrosion Resistance of Stainless Steels and Nickel Alloys A. John Sedriks, Office of Naval Research

Stainless Steels Stainless steels are corrosion-resistant iron- base alloys containing a maximum of 1.2% carbon and a minimum of 10.5% chromium by weight. This is the minimum amount of chromium that prevents the formation of rust in humid, unpolluted atmospheres, hence the designation “stainless.” The corrosion resistance of stainless steels is provided by a very thin and protective surface film, known as the passive film, which, when damaged, is selfhealing in the presence of a wide variety of environments. The Fe-Cr-Ni and Fe-Cr-Ni-Mn-N grades of stainless steels are austenitic and are popularly known by the former American Iron and Steel Institute (AISI) type numbers in the 300 and 200 series, respectively. The FeCr grades are martensitic at lower chromium levels, ferritic at higher chromium levels, and are known by numbers in the 400 series. The most popular austenitic, ferritic, and martensitic grades have been type 304 (containing 19% Cr, 10% Ni, and also known by the Unified Numbering System (UNS) number, S30400), type 430 (17% Cr, S43000), and type 410 (12% Cr, S41000), respectively. Another popular grade has been type 409 (11% Cr, S40900) because of its use in automobile exhaust systems. Duplex grades (containing approximately 50% austenite and 50% ferrite) and precipitation-hardening grades (mostly martensitic) are also available for higher strength applications. Stainless steels are used for consumer products; for machinery; in architecture; for military applications (particularly for nonmagnetic hulls of submarines and mine countermeasure vessels); and for equipment in the petroleum, chemical, aerospace, power, and process industries. Environmental initiatives, such as flue gas desulfurization in the power industry and the adoption of closed-loop (zero discharge) processes in the pulp and paper industry, have increased the demand for stainless steels. The manufacturers of stainless steels have tended to specialize as either flat-product producers or long-product producers. Flat products include plate, sheet, strip, and foil, whereas long products include bar, rod, wire, and forging billets. Products such as forgings, welded pipe and tube, seamless pipe and tube, fittings, and weld fillers are made from either long or flat products by mills dedicated to their production. Castings and powder metallurgy products are typically custom melted by specialty producers. Some 180 different alloys can be recognized as belonging to the stainless steel group and currently are produced, with alloy content adjusted to give improved resistance to pitting and crevice corrosion, intergranular corrosion caused by sensitization, stress-corrosion cracking (SCC) and hydrogen embrittlement, general corrosion, and attack by high-temperature gases (Fig. 1).

Fig. 1 Compositional and property linkages for stainless steels Pitting and Crevice Corrosion. Resistance to pitting and crevice corrosion usually is improved by alloying the austenitic and duplex grades further with chromium, molybdenum, and nitrogen, and the ferritic grades with chromium and molybdenum. The beneficial effects of these alloying elements are complex and interactive. Attempts have been made by suppliers of stainless steels and nickel-base alloys to develop a compositionally derived pitting- and crevice-corrosion-resistance index known as the pitting-resistance equivalent number (PREN). The PREN is given by the alloying-element parameter %Cr + 3.3%Mo + 16%N + 1.65%W. In general, the larger the numerical value of PREN is, the higher the pitting and crevice-corrosion resistance will be, although a high numerical value of PREN should not be viewed as an absolute guarantee of freedom from localized attack. The major drawback in using a parameter based only on alloy content is that it ignores the often-found detrimental effects of microstructural constituents such as manganese sulfide inclusions, sigma and chi phases, chromium depleted zones, and alloying element segregation due to coring produced by weld solidification. However, PREN provides some guidance for alloy selection for service in oxidizing chloride or acidic environments. Among the proprietary stainless steels with a large numerical value of PREN are the superaustenitics, such as 254SMO (S31254), AL6XN (N08367), 925hMo (N08925), and 654SMO (S32654); the superferritics, such as 29-4C (S44735), Sea-Cure (S44660), and Monit (S44635); and the superduplexes, such as DP-3W (S39274), Ferralium 255 (S32550), SAF 2507 (S32750), Zeron 100 (S32760), and Uranus 52N+ (S32520). Intergranular corrosion results from chromium depletion in the alloy matrix near grain- boundary chromium carbides (and sometimes nitrides). These can be precipitated during welding or some other high-temperature exposure. This depletion occurs at certain time-temperature combinations that are sufficient to precipitate chromium carbide but insufficient to rediffuse chromium back into the austenite near the carbide. For example, heating type 304 stainless steel containing 0.039% carbon for 10 h at 700 °C (1290 °F) results in the formation of chromium carbides that reduce the chromium level from 19% to less than 13% in the region next to the grain-boundary carbide precipitate, resulting in a loss of corrosion resistance in this region. This depletion is known as sensitization. Similar chromium depletion occurs in high-nitrogen duplex stainless steels due to the formation of chromium nitrides. Remedial measures for sensitization include alloying-element control procedures such as lowering the carbon content to 0.03% maximum, as in 304L (S30403), or stabilizing with titanium or niobium plus tantalum alloying additions (Fig. 1), or metallurgical treatments such as postweld

annealing to re-diffuse chromium back into the depleted regions. It should be noted that duplex stainless steels do not exhibit sensitization at the austenite-ferrite grain boundaries because of faster chromium diffusion and carbide growth kinetics in the ferrite phase. Nickel-base alloys are generally resistant to sensitization, because they are usually made either with sufficiently low carbon content or are stabilized with niobium. Stress-corrosion cracking may occur in the presence of a tensile stress and a specific corrodent. Corrodents known to cause SCC in stainless steels are chloride solutions at elevated temperatures, caustic solutions, acids, aqueous solutions containing sulfur compounds, and high-temperature (300 °C, or 570 °F) water containing traces of dissolved oxygen. For SCC caused by sensitization, the remedies are the same as for intergranular corrosion. In other cases, susceptibility to SCC can be minimized or eliminated by alloying additions, metallurgical treatments, mechanical treatments, and chemical/electrochemical treatments. For example, susceptibility to SCC can be minimized by increasing the nickel content of the alloy (for austenitics only), selecting a suitable ferritic or duplex stainless steel instead of an austenitic one, lowering the service temperature, stress-relief annealing, shot peening to introduce compressive surface stresses, adding chemical corrosion inhibitors to the service environment, and cathodic protection. The last procedure should not be used for martensitic or precipitation-hardening stainless steels, which crack by a hydrogen embrittlement mechanism. For martensitic or precipitation-hardening stainless steels, tempering heat treatments to produce lower strengths may improve cracking resistance. General corrosion (i.e., nonlocalized corrosion) can be encountered in strong acids or alkalies. Corrosion resistance in sulfuric and organic acids is promoted by alloying with copper, molybdenum, and nickel. Among alloys used in the chemical industry for sulfuric acid service are the copper-containing stainless steels and higher alloys, such as alloy 904L (N08904), 20Cb-3 (N08020), alloy 825 (N08825), and the cast stainless steel CN-7M (J95150), whereas higher-silicon stainless steel, such as S32615, has been used for handling hot, concentrated sulfuric acid. Phosphoric acid is handled by highmolybdenum grades, such as alloy 28 (N08028), G-30 (N06030), and 3127hMo (N08031). Nitric acid at most concentrations and temperatures can be handled by less highly alloyed stainless steels, such as type 304L (S30403), although higher- chromium stainless steels, such as type 310S (S31008) and alloy 800 (N08800), as well as silicon-containing stainless steels, such as S30600 and S30601, have been used for very hot, concentrated nitric acid. The ferritic stainless steel E-Brite (S44627) as well as the low-carbon version of commercially pure nickel, Nickel 201 (N02201), have been used to handle elevated- temperature caustic environments. High-Temperature Corrosion. The various types of attack by high-temperature gases usually are referred to as oxidation, sulfidation, carburization, nitriding, and halogen-gas corrosion. In oxidizing, sulfidizing, and carburizing gases, high chromium contents, such as in type 310 (25% Cr, S31000) or its cast variant HK (J94224), improve resistance to attack. In addition, alloying with aluminum and silicon can be beneficial to oxidation resistance, as in type 406 (3.5% Al) and in type 302B (S30215, 2.5% Si), respectively. Resistance to nitriding is improved by alloying with nickel, as in RA-330 (N08330). For stainless steels, the upper temperature limit for operation in dry chlorine is approximately 320 °C (600 °F), with the presence of water vapor accelerating corrosion.


Nickel-Base Alloys Nickel-base alloys make up an important segment of the corrosion-resistant materials, taking over from stainless steels and other alloys as service temperature or environment corrosivity increases. Commercially pure nickel, Nickel 200 (N02200), or its low-carbon version, Nickel 201 (N02201), is used as a corrosion-resistant material in food processing and in high-temperature caustic and gaseous chlorine or chloride environments.

However, alloying of nickel with other elements (e.g., chromium, copper, or molybdenum) greatly broadens its use in corrosion- resistant applications (Fig. 2).

Fig. 2 Compositional and property linkages for nickel-base alloys Nickel-Chromium Alloys. By far, the largest family of nickel alloys is that based on the nickel-chromium system, with alloy 600 (N06600) being the prototype (Fig. 2). Chromium imparts resistance to oxidizing environments and high-temperature strength. Increasing chromium to 30%, as in alloy 690 (N06690), also increases resistance to SCC in high-temperature (300 °C, or 570 °F) water and to corrosion in nitric acid solutions, steam, oxidizing gases, and shipboard waste-incinerator environments. Increasing chromium to 50%, as in IN- 657 (N07765), increases resistance to melting sulfates and vanadates found in fuel ash. High-temperature oxidation resistance of nickel-chromium alloys is improved further by alloying with aluminum, as in alloys 601 (N06601) and 617 (N06617). Alloying additions of silicon and rare earth elements (e.g., cerium, yttrium, and lanthanum) also increase oxidation resistance. Among alloys that take advantage of the benefits of rare earth element additions on oxidation resistance are the silicon-containing alloys, such as 353MA (S35315) and 45TM (N06045), and the aluminum-containing alloys, such as 602CA (N06025) and 214 (N07214). Of importance for use in aqueous reducing acids, oxidizing chloride solutions, and seawater in the presence of crevices and tight joints are the Ni-Cr-Mo alloys, such as 625 (N06625), C-276 (N10276), C-22 (N06022), 59 (N06059), 686 (N06686), and C-2000 (N06200). For these alloys to exhibit the maximum resistance to crevice corrosion in seawater environments, they should be free of deleterious precipitate phases and chromiumdepleted surface layers. It is important, therefore, to remove by pickling, electropolishing, or mechanical processes, such as abrasion or grinding, any surface layers that have become depleted in chromium during hightemperature (above 980 °C, or 1800 °F) manufacturing steps. Low-level titanium and aluminum alloying additions to nickel-chromium alloys and to Ni- Cr-Mo alloys result in strengthening by the precipitation of the γ′ phase, without loss of corrosion resistance, as in alloys X-750 (N07750) and 725 (N07725), respectively. Cobalt and other alloying additions provide to jet engine materials (superalloys) a combination of hightemperature strength and creep resistance, with oxidation and sulfidation resistance. Oxide dispersion strengthening, in addition to γ′ strengthening, is used in the mechanically alloyed materials MA 754 (N07754) to provide high-temperature strength and oxidation resistance at the very high temperatures (approximately 1200 °C, or 2200 °F) encountered in molten glass processing and in reheating furnaces used in steel production.

Nickel-molybdenum alloys, such as B-2 (N10665), have excellent corrosion resistance in hydrochloric acid with low oxidizer content, whereas nickel-silicon alloys, such as D-207, have good corrosion resistance in hot, concentrated sulfuric acid. Another technologically important group of materials are the lower-nickel (30 to 40%) Ni- Cr-Fe alloys that were originally developed to conserve nickel. The prototype, alloy 800 (N08800), is a general-purpose alloy with good high-temperature strength and good corrosion resistance in steam and in oxidizing or carburizing gases. Further alloying with molybdenum and copper, as in alloys 825 (N08825), G-3 (N06985), G-30 (N06030), 28 (N08028), 20Cb- 3 (N08020), and 20Mo-4 (N08024), improves resistance to localized corrosion in chlorides and resistance to general corrosion in reducing acids. Nickel alloys exhibit high resistance to corrosive attack under nitriding conditions (e.g., in dissociated ammonia) and in chlorine or chloride gases. Corrosion in the latter at elevated temperatures proceeds by the formation and volatilization of chloride scales. High nickel contents in the alloys are beneficial, because nickel forms one of the least volatile chlorides. Conversely, in sulfidizing environments, high-nickel alloys without chromium can undergo corrosive attack due to the formation of a low-melting-point eutectic. Nickel-copper alloys represent another technologically important group of materials. At higher nickel contents are alloys 400 (N04400), R-405 (N04405), and K-500 (N05500), which are used for certain corrosive chemicals, such as hydrofluoric acid, and for seawater. At higher copper contents are the cupronickels, such as 706 (C70600) and 715 (C71500), which are widely used for seawater applications because of their corrosion resistance and the ease with which marine fouling can be removed by mechanical processes.



“Corrosion Engineering Bulletin 1–6,” International Nickel Company, Inc., 1962–1980 J.R. Davis, Ed., ASM Specialty Handbook: Stainless Steels, ASM International, 1994 C.P. Dillon, Corrosion Resistance of Stainless Steels, Dekker, 1995 M.G. Fontana and N.D. Greene, Corrosion Engineering, 2nd ed., McGraw-Hill, 1979 W.Z. Friend, Corrosion of Nickel and Nickel- Base Alloys, Wiley, 1979 F.L. LaQue and H.R. Copson, Ed., Corrosion Resistance of Metals and Alloys, 2nd ed., Reinhold, 1963 D. Peckner and I.M. Bernstein, Ed., Handbook of Stainless Steels, McGraw-Hill, 1977 A.J. Sedriks, Corrosion of Stainless Steels, 2nd ed., Wiley, 1996

S. Yu, Corrosion Resistance of Titanium Alloys, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 703–711

Corrosion Resistance of Titanium Alloys Steven Yu, 3M Company

Introduction TITANIUM AND ITS ALLOYS are valuable engineering materials due to their combination of light weight, high strength-density ratio, and stability in a wide range of environmental conditions (Ref 1, 2, 3, 4, 5, 6). Yet, historically high cost per weight has limited the use of titanium and titanium-base alloys to mission-critical engineering applications such as aerospace, military, or biomedical industries. Recent availability of commercial-grade titanium from the Russian Federation and the People's Republic of China has made the application of titanium-base alloys more accessible for the general consumer industry. Information on the engineering applications of titanium, its mechanical properties, and its alloy variants is readily available in literature (Ref 1, 2, 3, 4, 5, 6). The scope of this article is to provide the reader with a background in the complex relationship between titanium and its alloys with aqueous environments, which is dictated by the presence of a passivating oxide film.

References cited in this section 1. M.J. Donachie, Jr., Titanium: A Technical Guide, 2nd ed., ASM International, 2000 2. C.R. Brooks, Heat Treatment, Structure and Properties of Nonferrous Alloys, American Society for Metals, 1982 3. R. Boyer, G. Welsch, and E.W. Collings, Ed., Materials Properties Handbook: Titanium Alloys, ASM International, 1994 4. T.P. Hoar and D.C. Mears, Corrosion-Resistant Alloys in Chloride Solutions: Materials for Surgical Implants, Proc. R. Soc. (London) A, Vol 294 (No. 1438), 1966, p 486–510 5. A. McQuillan, Titanium Science and Technology, Plenum Press, 1973, p 915–922 6. R. Schutz, An Overview of Beta Titanium Alloy Environmental Behavior, Beta Titanium Alloys in the 1990's, D. Eylon, R. Boyer, and D. Koss, Ed., Minerals, Metals and Materials Society, 1993, p 75–91

S. Yu, Corrosion Resistance of Titanium Alloys, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 703–711 Corrosion Resistance of Titanium Alloys Steven Yu, 3M Company

Corrosion Resistance of Titanium and Titanium-Base Alloys in Aqueous Environments

Pure titanium is highly corrosion resistant in aqueous environments that encompass any pH > 2, even those that contain aggressive anionic species (including Cl-) (Ref 6, 7). Nevertheless, metallic titanium is thermodynamically reactive, as indicated by its relatively negative reversible potential on the electromotive force (emf) scale (E0 = -1.63 VNHE). (NHE is normal hydrogen electrode.) As a result of its reactivity, metallic titanium readily oxidizes during exposure to air as well as during exposure to aqueous and nonaqueous electrolytes (Ref 6, 8, 9, 10, 11, 12, 13, 14, 15). Such oxidation leads to the formation of titanium-base oxides, hydrated complexes, or aqueous cationic species as a result of active anodic dissolution. The oxide and hydrated-complex layers function as barriers between the surrounding environment and the underlying metallic titanium, which inhibit the subsequent oxidation of metallic titanium across the metal/barrier layer/solution interface. Therefore, further oxidation of titanium can only occur by anion and cation movement across the oxide through diffusion and migration by potential field-assisted movement through this barrier (Ref 15, 16, 17, 18). Coherent titanium oxide films resist uniform corrosion in many environments, such as all natural waters, including distilled, fresh, and seawater (aerated and deaerated), as well as brine solutions, to temperatures in excess of 200 °C (390 °F) (Ref 6). Furthermore, titanium is generally corrosion resistant in the presence of oxidizing-acid media. Oxidizing acids, including chromic, nitric, perchloric, and hypochloric acids, readily oxidize titanium to form thermodynamically stable TiO2 where oxide formation and further oxide thickening promotes additional passivity. In contrast, passivated titanium exhibits poor resistance to corrosion in reducingacid environments, such as hydrochloric or sulfuric acids, where the passivating TiO2 can be reduced to a soluble form of oxidized titanium. Therefore, alloying additions that can promote the stability of protective titanium oxide films in reducing-acid environments are beneficial in inhibiting corrosion of titanium-base alloys.

References cited in this section 6. R. Schutz, An Overview of Beta Titanium Alloy Environmental Behavior, Beta Titanium Alloys in the 1990's, D. Eylon, R. Boyer, and D. Koss, Ed., Minerals, Metals and Materials Society, 1993, p 75–91 7. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, 1st ed., NACE, 1974 8. R.W. Schutz and J.S. Grauman, “Compositional Effects on Titanium Alloy Repassivation Potential in Chloride Media,” Second International Conference on Localized Corrosion, (Orlando, FL) NACE, 1987 9. P.Y. Park et al., The Corrosion Behavior of Sputter-Deposited Mo-Ti Alloys in Concentrated Hydrochloric Acid, Corros. Sci., Vol 38 (No. 10), 1996 p 1649–1667 10. V. Andreeva et al., Investigation of the Corrosion Resistance and the Electrochemical and Mechanical Properties of Titanium-Niobium Alloys, Corrosion of Metals and Alloys, N. Tomashov and E. Nirolybev, Ed., Israel Program for Scientific Translation, Jerusalem, 1966, p 34–47 11. P.J. Bania, Beta-Titanium Alloys and Their Role in Titanium Industry, D. Eylon, R. Boyer, and D. Koss, Ed., TMS, 1993, p 4 12. M. Levy and G. Sklover, Anodic Polarization of Titanium and Titanium Alloys in Hydrochloric Acid, J. Electrochem. Soc., Vol 116 (No. 3), 1969, p 323–328 13. M. Itagaki, R. Oltra, and B. Vuillemin, Quantitative Analysis of Iron Dissolution During Repassivation of Freshly Generated Metallic Surfaces, J. Electrochem. Soc., Vol 144 (No. 1), 1997, p 64–72 14. A. Mazhar, F. Heakal, and A. Gad-Allah, Anodic Behavior of Titanium in Aqueous Media, Corros. Sci., Vol 44, 1988, p 705–710 15. D. Tomashov et al., The Passivation of Alloys on Titanium Bases, Electrochim. Acta, Vol 19, 1974, p 159–172

16. P.A. Mausli, S.G. Steinemann, and J.P. Simpson, “Properties of Surface Oxides on Titanium and Some Titanium Alloys,” Proceedings of the Sixth World Conference on Titanium (France), P. LaCombe, R. Tricot, and G. Bèranger, Ed., Sociètè Francaise de Métallurgie, 1988 17. N. Khalil and J.S.L. Leach, The Anodic Oxidation of Valve Metals, Part I: Determination of Ionic Transport Numbers by Alpha Spectrometry, Electrochim. Acta, Vol 31 (No. 10), 1986, p 1279–1285 18. J. Leach and B. Pearson, The Effect of Foreign Ions upon the Electrical Characteristics of Anodic ZrO2 Films, Electrochim. Acta, Vol 29 (No. 9), 1984, p 1271–1282


Passivating Titanium Oxides Titanium is one of the thermodynamically reactive metals generally described as valve metals due to the ability of these metals to allow relatively fast cathodic reaction rates while limiting the reaction rate of the anodic reactions (Ref 19). Limited anodic reaction rates result from the formation of thermodynamically or kinetically stable oxide films. Such oxide films reduce reactivity and enhance corrosion resistance of titanium within a wide range of pH, as shown in Fig. 1. Moreover, titanium oxides, formed near the respective open-circuit potentials (OCPs) of titanium, have been generally agreed on to be amorphous (Ref 20, 21, 22). The amorphous nature of titanium oxide is attributed to a cation egress mechanism due to an applied potential field gradient, which results in the formation of the oxide primarily at the solution-oxide interface (Ref 22, 23). Therefore, there is limited epitaxial constraint to form a crystalline oxide. However, titanium oxides can undergo a phase transformation from amorphous to one of the following crystalline structures during anodic polarization or during heat treatments: (a) anatase (tetragonal), (b) rutile (tetragonal), or (c) brookite (orthorhombic) (Ref 22, 24). The lattice parameters of TiO2 are presented in Table 1.

Fig. 1 Potential-pH equilibrium diagram for the titanium-water system at 25 °C (77 °F). The diagram was established by considering, as derivatives of the tri- and tetravalent titanium, the hydroxide Ti(OH)3 and the hydrated oxide TiO2-H2O. Lines a and b establish the stability region of water. Consult Ref 7 for further explanation. Table 1 Lattice parameters of crystalline titanium oxide Crystal structure TiO2-(anatase)

Lattice constant a, nm Lattice constant b, nm Lattice constant c, nm … 0.937 tetragonal 0.3733

0.4584 TiO2-(rutile) tetragonal TiO2-(brookite) orthorhombic 0.5436

… 0.9166

0.2953 0.5135

The morphology of anodically formed oxides on valve metals, specifically titanium, has been described as a dual-layered oxide structure (Ref 14, 15, 25). The dual-layered oxide structure is composed of the following: (a) an inner compact and protective oxide at the metal-oxide interface, and (b) a porous outer oxide structure. The inner compact layer is primarily responsible for the passivation behavior of the oxide and is also responsible for the majority of the potential drop across the oxide film. The outer layer is a defected and porous oxide that has limited passivation characteristics (Ref 15). However, anodic polarization to more noble potentials within the passive potential range of titanium results in the thickening of the compact protective oxide by the following: (a) growth of the thin and protective oxide by metal cation egress, and/or (b) the conversion of the outer oxide into the more protective oxide (Ref 15). Alloying additions, which can improve the efficiency of passivation, enhance the resistance of the passivating films to chemical dissolution, and/or inhibit the direct electrochemical dissolution of titanium through the oxide, should reduce passive current densities. This becomes crucial for cases of non-steady-state or rapid oxide formation, during which a majority of the anodic charge (~95%) is accounted for by the introduction of oxidized titanium into solution through the direct anodic dissolution of titanium (Ref 15, 26).

References cited in this section 7. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, 1st ed., NACE, 1974 14. A. Mazhar, F. Heakal, and A. Gad-Allah, Anodic Behavior of Titanium in Aqueous Media, Corros. Sci., Vol 44, 1988, p 705–710 15. D. Tomashov et al., The Passivation of Alloys on Titanium Bases, Electrochim. Acta, Vol 19, 1974, p 159–172 19. M.J. Chappell and J.S.L. Leach, Passivity and Breakdown of Passivity of Valve Metals, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 1003–1034 20. E.J. Kelly and H.R. Bronstein, Kinetics and Mechanism of the Hydrogen Evolution Reaction of Titanium in Acidic Media, J. Electrochem. Soc., Vol 131 (No. 10), 1984, p 2232–2238 21. N. Khalil, A. Bowen, and J.S.L. Leach, The Anodic Oxidation of Valve Metals, Part II: The Influence of the Anoidizing Conditions on the Transport Processes During the Anodic Oxidation of Zirconium, Electrochim. Acta, Vol 33 (No. 12), 1988, p 1721–1727 22. T. Shibata and Y. Zhu, The Effect of Film Formation Conditions of the Structure and Composition of Anodic Oxide Films on Titanium, Corros. Sci., Vol 37 (No. 2), 1995, p 253–270

23. J.A. Davies et al., The Migration of Metal and Oxygen During Anodic Film Formation, J. Electrochem. Soc., Vol 112 (No. 7), 1965, p 675–680 24. G.V. Samsonov, The Oxide Handbook, 2nd ed., IFI/Plenum, 1982 25. J. Pan, D. Thierry, and C. Leygraf, Electrochemical Impedance Spectroscopy Study of the Passive Oxide Film on Titanium for Implant Applications, Electrochim. Acta, Vol 41 (No. 7/8), 1996, p 1143– 1153 26. D.G. Kolman and J.R. Scully, Electrochemistry and Passivty of a Ti-15Mo-3Nb-3Al Beta-Titanium Alloy in Ambient Temperature Aqueous Chloride Solutions, J. Electrochem. Soc., Vol 140 (No. 10), 1993, p 2771


Corrosion Vulnerability of Titanium and Titanium Oxides: The Effect of Selected Environments The presence of a coherent air-formed or an anodically formed titanium oxide surface film inhibits corrosion and reduces active dissolution rates of metallic titanium. However, titanium oxide films are susceptible to failures that lead to accelerated mass loss rates. Oxide failure mechanisms can be classified in the following categories: • • •

Spatially localized oxide film breakdown by the ingress of aggressive anions, such as chlorides (Cl -) and bromides (Br-) Spatially local or homogenous chemical dissolution of the oxide in a strong reducing-acid environment Mechanical disruptions or depassivation, such as scratching, abrading, or fretting

Aggressive Anions. The presence of TiO2 is a formidable barrier to uniform corrosion, but it can fail and lead to localized corrosion, including pitting, in the presence of aggressive anion species. Aggressive anion species, especially halide ions such as Cl-, cause pitting (Ref 6, 22, 27, 28, 29). A possible mechanism of TiO2 failure, which subsequently leads to pitting, is anion migration to the Ti/TiO2 interface to form Ti/Ti-X/ TiO2, where X is the aggressive anion species. According to one theory, the internal stresses associated with the Ti-X formation lead to rupture of the covering oxide film (Ref 19, 30). However, pitting of passivated titanium is limited to high anodic overpotentials, Eapp > 5 V above OCP, even in aggressive halide-containing solutions, as indicated in Table 2. It is evident that titanium is highly resistant to pitting in solutions containing aggressive anion species, including Cl- and Br-, at temperatures of 25 °C to 100 °C (77 to 212 °F) and pH > 0.

Table 2 Pitting potential of commercially pure titanium Environment

Temperature Pitting potential Source °C °F 30 86 +6 VSCE Ref 31 1 M NaBr 25 77 +8.5 VSCE Ref 30 5% NaCl-HCl, pH = -0.05 +6 VSCE Ref 6 5% NaCl-HCl, pH = 3.5 Boiling 25 77 +14 VSCE Ref 33 1 M KCl Potential reference is saturated calomel electrode (SCE). In addition to a resistance to pitting, titanium exhibits resistance to crevice corrosion that is a result of: (a) deaeration, (b) separation of anode and cathode, (c) metal-ion hydrolysis, (d) localized acidification, and (e) a breakdown mechanism. As a general rule, titanium is not susceptible to crevice corrosion at T < 70 °C (158 °F), regardless of the solution pH or chemistry, as shown in Fig. 2 (Ref 6, 34). At T > 70 °C (158 °F), initiation of crevice corrosion on titanium can be attributed to crystallization of titanium oxide, and increased passive dissolution, which lead to the five-stage process discussed previously (Ref 34). The anodically formed titanium oxide is amorphous at T < 60 °C (140 °F). The phase transformation of amorphous titanium oxide to a crystalline oxide can occur by anodic polarization (Ref 15, 22, 35). However, at T > 60 °C (140 °F), compressive stresses within the oxide can cause crystallization of the oxide. The resulting crystalline oxide film contains defects such as grain boundaries and intersections of screw dislocations with free surfaces, which can act as preferential sites for oxide failure or dissolution (Ref 9, 13, 36). The formation of grain boundaries, which can act as ionic diffusion pathways, allows for anion transport into the oxide (Ref 34). Moreover, increased passive dissolution at elevated temperatures can lead to metal-ion hydrolysis that generates acidity in an occluded geometry (Ref 22). Consequently, titanium with crystalline oxide films can be susceptible to crevice corrosion as well as pitting corrosion. Nevertheless, resistance to crevice corrosion at T < 70 °C (158 °F) does not preclude the possibility of a localized solution acidification mechanism leading to accelerated passive dissolution, because anodic dissolution rates of active and passive titanium are inversely proportional to related solution pH (Ref 12). The topic of local acidification is addressed subsequently. Note that alloying additions of molybdenum, zirconium, and palladium to titanium do raise the critical temperature of 70 °C (158 °F), as shown in Fig. 2.

Fig. 2 Approximate temperature limits for crevice corrosion resistance of titanium alloys by group in various chloride brines. Group A: commercially pure titanium (grade 2) and beta titanium alloys. Group B: Beta-C (Ti-3%Al-8%V-6%Cr-4%Zr-4%Mo), Transage-207 (Ti-8%Mo-2.5%Al-9%Zr-2%Sn), and Ti-8-8-2-3 (Ti-8%Mo-8%V-2%Fe-3%Al). Group C: Beta-C/Pd (Ti-3%Al-8%V-6%Cr-4%Zr-4%Mo0.05%Pd), Beta-21S (Ti-15%Mo-2.7%Nb-3%Al-0.2%Si), Ti-15-5 (Ti-15%Mo-5%Zr), and Beta III (Ti-

11.5%Mo-6%Zr-4.5%Sn). It is evident that the alloying additions of molybdenum, zirconium, and palladium were beneficial in raising the 70 °C (158 °F) threshold for crevice corrosion. Source: Ref 6 Reducing Acids. Titanium is generally corrosion resistant in the presence of oxidizing-acid media but exhibits performance limitations in the presence of reducing-acid media. Oxidizing acids, such as chromic, nitric, perchloric, and hypochloric acids, can oxidize titanium to form TiO2, where further oxide thickening promotes passivity. In contrast, passivated titanium may undergo activation through the dissolution of TiO2 in a reducing environment. As Ti4+ is reduced to a more soluble Ti3+, passive film dissolution can occur, as previously discussed. The dissolution of titanium in the form of Ti3+ can occur in reducing acids such as hydrochloric acid. Interestingly, alloying additions can improve the corrosion resistance of titanium-base alloys in such an environment (Ref 6, 15). Some common acids, which are classified as reducing acids, are hydrochloric, sulfuric, phosphoric, and hydrofluoric acids. The breakdown of anodic or air-formed titanium oxide has been correlated with the conversion of TiO 2 to hydrated TiOOH and the presence of Ti3+ (Ref 37, 38). It has been proposed (Ref 37) that the formation of Ti3+ in the form of TiOOH or its hydrated form of TiOOH · H2O, for example, Ti(OH)3, through the reduction of Ti4+ in TiO2 can occur at the oxide-solution interface. This can result in oxide film failure, because TiOOH · H2O should readily dissolve in acidic solutions. The potential-pH equilibrium diagram for the Ti-H2O system should be modified to include TiH2 (Ref 39). In such a case, an equilibrium involving only the hydrated form of trivalent titanium may occur between the regions of TiO2 and TiH2 equilibrium (Ref 37). The formation of Ti3+ is critical, due to the higher solubility of Ti3+ at low pH in comparison to Ti4+, as shown in Fig. 1. The solubility of Ti3+ accounts for active anodic dissolution of titanium in reducing acids. Mechanical Attack. Mechanical disruption of TiO2 by scratching, abrading, or fretting can also result in the exposure of the underlying thermodynamically reactive metallic titanium to an aqueous environment. However, repassivation can occur to reform an oxide layer to isolate the underlying reactive metal from the aqueous environment. When repassivation occurs, the high current densities associated with repassivation do not all go into the formation of passive oxides. Current densities have been recorded as high as 100 A/cm2 (645 A/in.2) for thin-film fracture tests of metallic titanium in deaerated 5 M HCl at 25 °C (77 °F) (Ref 40). However, it was recorded that the overall current efficiency for oxide formation on decay to steady state was less than 10% (Ref 40). Therefore, a majority of the repassivation charge went into metallic titanium dissolution, either as Ti3+ or TiO2+ (Ref 40, 41). Consequently, there can be a significant introduction of dissolved metallic ionic species into the local solution chemistry during repassivation. Within a tight geometry, such influx of titanium ions or other metal ions can lead to localized solution acidification by metal-ion hydrolysis. The low current efficiency associated with repassivation contrasts the near 100% charge efficiency associated with the slower passive film growth during anodic potentiostatic polarization of titanium in 0.5 M H2SO4 at temperatures ranging from 303 to 353 K (86 to 176 °F) (Ref 22). Therefore, reactivation current efficiency during repassivation and film growth during exposure to reducing acids may depend critically on the rate of oxidation and the disordered state of the oxide. The release of titanium ions into the surrounding aqueous environment, as a result of metallic dissolution, is a critical issue with respect to corrosion in reducing acids. Introduction of titanium ions into the surrounding aqueous environment can lead to solution acidification by metallic-ion hydrolysis (Ref 42). It was shown that the dominant dissolved titanium species resulting from scratch-repassivation of titanium disks was Ti3+, which was reported as being hydrolyzable (Ref 41, 42). In addition, it has been shown that the production of Ti3+ occurs over the duration of repassivation, due to the low overall current efficiency of TiO2 formation immediately after rapid mechanical depassivation of microelectrodes that enabled high dissolution rates (Ref 40). Trivalent titanium ions (Ti3+) can undergo metal-ion hydrolysis primarily by the following reaction: Ti3+ + H2O → TiOH2+ + H+

(Eq 1) 3+

The resulting hydrolysis of Ti can produce a solution with a pH as low as 0, as shown when 1 M TiCl3 is added to deaerated water (Ref 40). Indeed, the pH near a crack tip in Ti-8%Al- 1%Mo-1%V has been reported to be as low as 1.7 (Ref 43). Also, a pH < 1.3 was reported near a corroding pit on titanium that was exposed to neutral chloride solution (Ref 41). The decrease in localized solution pH due to hydrolysis can lead to enhanced titanium dissolution in reducing acid that, in turn, further acidifies the surrounding solution and causes even more dissolution of titanium, because the active and passive dissolution rates for titanium are well known to be functions of pH, as shown in Fig. 3 (Ref 26, 41, 44).

Fig. 3 Electrochemical polarization tests of titanium and titanium-base alloys, including Ti-6%Al-4%V (Ti-6-4), Ti-15%Mo-3%Nb-3%Al (Ti-15-3-3), Ti-13%Nb-13%Zr (Ti-13-13), Ti-55%Ni, and commercially pure (CP) titanium (grade 2), revealed that titanium and titanium-base alloys are spontaneously passive in a deaerated 0.1 M NaCl solution. The potentiodynamic scans were conducted at a scan rate of 0.1 mV/s. SCE, saturated calomel electrode The resulting critical corrosion-related issues for titanium alloys include the possibility of: (a) occluded site deaeration, (b) partial or complete separation of anode and cathode sites, and (c) the metal dissolution/hydrolytic acidification/chloride ion (Cl-) migration mechanism promoting a local buildup of acidity and high Cl- concentrations adjacent to the alloy structures. In titanium applications that combine abrasive actions in a crevice, local acidification may be accelerated by the film rupture and repassivation mechanisms. The issues of local solution acidification and Cl- accumulation do point toward crevice corrosion. However, crevice corrosion, which is defined by local depassivation and active dissolution in the absence of active abrasion, is not expected of titanium unless it is exposed to low- pH solutions at temperatures above 70 °C (158 °F), (Ref 1, 6, 34, 45, 46, 47, 48). It is yet unknown whether abrasion in crevice geometry could activate crevice corrosion at T < 70 °C (158 °F).

References cited in this section 1. M.J. Donachie, Jr., Titanium: A Technical Guide, 2nd ed., ASM International, 2000 6. R. Schutz, An Overview of Beta Titanium Alloy Environmental Behavior, Beta Titanium Alloys in the 1990's, D. Eylon, R. Boyer, and D. Koss, Ed., Minerals, Metals and Materials Society, 1993, p 75–91 9. P.Y. Park et al., The Corrosion Behavior of Sputter-Deposited Mo-Ti Alloys in Concentrated Hydrochloric Acid, Corros. Sci., Vol 38 (No. 10), 1996 p 1649–1667 12. M. Levy and G. Sklover, Anodic Polarization of Titanium and Titanium Alloys in Hydrochloric Acid, J. Electrochem. Soc., Vol 116 (No. 3), 1969, p 323–328 13. M. Itagaki, R. Oltra, and B. Vuillemin, Quantitative Analysis of Iron Dissolution During Repassivation of Freshly Generated Metallic Surfaces, J. Electrochem. Soc., Vol 144 (No. 1), 1997, p 64–72 15. D. Tomashov et al., The Passivation of Alloys on Titanium Bases, Electrochim. Acta, Vol 19, 1974, p 159–172

19. M.J. Chappell and J.S.L. Leach, Passivity and Breakdown of Passivity of Valve Metals, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 1003–1034 22. T. Shibata and Y. Zhu, The Effect of Film Formation Conditions of the Structure and Composition of Anodic Oxide Films on Titanium, Corros. Sci., Vol 37 (No. 2), 1995, p 253–270 26. D.G. Kolman and J.R. Scully, Electrochemistry and Passivty of a Ti-15Mo-3Nb-3Al Beta-Titanium Alloy in Ambient Temperature Aqueous Chloride Solutions, J. Electrochem. Soc., Vol 140 (No. 10), 1993, p 2771 27. N. Casillas et al., Pitting Corrosion of Titanium, J. Electrochem. Soc., Vol 141 (No. 3), 1994, p 636–642 28. P. McKay and D. Mitton, An Electrochemical Investigation of Localized Corrosion on Titanium in Chloride Environments, Corrosion, Vol 41 (No. 1), 1985, p 52–61 29. G. Burstein and R. Souto, Observations of Localized Instability of Passive Titanium in Chloride Solution, Electrochim. Acta, Vol 40 (No. 12), 1995, p 1881–1888 30. J. Yahalom and A. Poznansky, Surface Energy and the Stability of the Passive Layer, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 328–336 31. T. Shibata and Y. Zhu, The Effect of Film Formation Potential on the Stochastic Processes of Pit Generation on Anodized Titanium, Corros. Sci., Vol 36(No. 1), 1994, p 153–163 33. I. Dugdale and J. Cotton, The Anodic Polarization of Titanium in Halide Solutions, Corros. Sci., Vol 4, 1964, p 397–411 34. J.J. Noel et al., Passive Oxide Film on Titanium in Aqueous Chloride Solution Probed by Electrochemistry, XPS, and In- Situ Neutron Reflectometry, Surface Oxide Films, The Electrochemical Society, 1996 35. R.D. Armstrong and R.E. Firman, Impedance of Titanium in the Active-Passive Transition, J. Electroanal. Chem. Interfacial Electrochem., Vol 34, 1972, p 391–397 36. J.R. Galvele, Present State of Understanding of the Breakdown of Passivity and Repassivation, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 285–323 37. C.K. Dyer and J.S.L. Leach, Reversible Reactions Within Anodic Oxide Films on Titanium Electrodes, Electrochim. Acta, Vol 23, 1978, p 1387–1394 38. D.J. Blackwood and L.M. Peter, Stability and Open Circuit Breakdown of the Passive Oxide Film on Titanium, Electrochim. Acta, Vol 33 (No. 8), 1988, p 1143–1149 39. T.R. Beck, Electrochemistry of Freshly- Generated Titanium Surfaces, Part I: Scraped Rotating Disk Experiment, Electrochim. Acta, Vol 18, 1973, p 807 40. D. Kolman and J. Scully, On the Repassivation Behavior of High Purity Titanium and Selected Alpha, Beta, Alpha + Beta Titanium Alloys in Aqueous Chloride Solutions, J. Electrochem. Soc., Vol 143 (No. 6), 1996, p 1847–1859 41. T.R. Beck, Localized Corrosion, B.F. Brown, Ed., NACE, 1974, p 644 42. C. Baes and R. Mesmer, The Hydrolysis of Cations, John Wiley & Sons, 1976, p 147–168

43. B.F. Brown, C.T. Fujii, and E.F. Dahlberg, J. Electrochem. Soc., Vol 116 (No. 218), 1969, p 218 44. M. Stern and H. Wissenberg, The Electrochemical Behavior and Passivity of Titanium, J. Electrochem. Soc., Vol 106 (No. 9), 1959, p 755–759 45. P. Mckay, Crevice-Corrosion Kinetics on Titanium and a Ti-Ni-Mo Alloy in Chloride Solutions at Elevated Temperatures, Corrosion Chemistry Within Pits, Crevices, and Cracks, HMSO Books, 1987 46. R.W. Schutz, “Understanding and Preventing Crevice Corrosion,” Corrosion 91, (Cincinnati, OH), NACE, 1991 47. L.A. Yao and F.X. Gan, Microelectrode Monitoring the Crevice Corrosion in Titanium, Corrosion, Vol 47 (No. 6), 1991, p 420–423 48. H. Shizhong and M. Xiaoxiong, Localized Corrosion in Titanium in Marine Environments, Corrosion and Corrosion Control for Offshore and Marine Construction, Pergamon Press, 1988


Effects of Alloying on Active Anodic Corrosion of Titanium Titanium alloys can be classified into three primary groups: (a) α titanium alloys with hexagonal close-packed (hcp) crystallographic structure; (b) β titanium alloys with body-centered cubic (bcc) crystallographic structures, including metastable β alloys; (c) and α+β titanium alloys, including near-α and near-β titanium alloys. Details regarding the alloy families are described in Ref 2 and 3. For titanium- base alloys containing β stabilizing elements, a molybdenum equivalency has been determined as:

(Eq 2)

where approximately 10% Mo equivalency is sufficient to stabilize completely the beta phase on quenching to room temperature from above the beta-transus temperature. Active Anodic Polarization Behavior of Titanium and Titanium-Base Alloys. Electrochemical polarization tests of titanium and titanium-base alloys, including Ti-6%Al-4%V (Ti-6-4), Ti-15%Mo-3%Nb-3%Al (Ti-15-3-3), Ti-13%Nb-13%Zr (Ti-13-13), Ti-55%Ni, and Ti-50%Cr, showed that titanium and the tested titanium-base alloys are spontaneously passive in a deaerated 0.1 M NaCl solution, as shown in Fig. 3. This is consistent with the stability of dominant TiO2 at these OCP values and pH, as shown in Fig. 1. From OCP to +2.5 VSCE, the tested alloys exhibited similar passive current densities of ~5 × 10-7 A/cm2 (3.2 × 10-6 A/ in.2). The corrected critical current densities (icrit) of the solution-treated alloys Ti-13-13, Ti-15-3-3, as well as commercially pure (CP) titanium (grade 2) were compared as functions of their respective molybdenum equivalency, as shown in Fig. 4. The corrected icrit was defined as the following: Corrected icrit = icrit(measured) + icathodic

(Eq 3)

where the icathodic can be determined by Tafel extrapolation to the Eapp, where the apparent icrit is observed. The icrit determined from the anodic polarization data, shown in Fig. 4, has been corrected to account for the cathodic current densities. Of the three alloys, solution-treated (ST) Ti-15-3-3, a metastable beta alloy, exhibited the lowest icrit in 5 M HCl. Moreover, ST Ti-15-3-3 did not exhibit an active/passive transition in the 5 M HCl, as shown in Fig. 4. Instead, spontaneous passivity was observed at the OCP of Ti- 15-3-3. Researchers also reported that Ti-15-3-3 was spontaneously passive at its OCP in 5 M HCl at room temperature (Ref 40).

Fig. 4 Electrochemical polarization tests of titanium and titanium-base alloys, including Ti-6%Al-4%V (Ti-6-4), Ti-15%Mo-3%Nb-3%Al (Ti-15-3-3), Ti-13%Nb-13%Zr (Ti-13-13), Ti-55%Ni, and commercially pure (CP) titanium (grade 2), in a deaerated 5 M HCl solution. The potentiodynamic scans were conducted at a scan rate of 0.1 mV/s. SCE, saturated calomel electrode Depassivation and Activation: Effects of Alloying Additions on Titanium at OCP. The dissolution of the respective native oxides and the subsequent surface activation of CP titanium as well as binary alloys, including Ti-45%Nb and Ti-50%Zr with native air-formed oxides (~2 to 3 nm in thickness) in 5 M HCl, were indicated by decreasing OCP. The decreasing OCP was followed by a sudden potential drop, as shown in Fig. 5 and 6. In 5 M HCl, the OCPs of CP titanium, Ti-45%Nb, and Ti-50%Zr were within the thermodynamically stable region of Ti3+ after the 12 h immersion, that is, after the potential drop, as revealed by the overlay of the OCPs onto the revised Pourbaix diagram of titanium at 37 °C (99 °F), as shown in Fig. 7. Similar decreases and stepfunction drops in OCPs of titanium were recorded in sulfuric and hydrochloric acids (Ref 38, 49, 50). It was confirmed that the OCP breakdown of the passive film on titanium was by chemical dissolution (Ref 38). The potential drops were not observed in deaerated 0.1 M NaCl (pH 6.8) for CP titanium and the titanium-base alloys (Fig. 5, 6). In this near-neutral pH solution, the air-formed oxides of all these materials were thermodynamically stable and resisted chemical dissolution, in contrast to what was experienced in deaerated 5 M HCl for the titanium-base materials (Fig. 5, 6). An overlay of the OCPs of CP titanium, Ti-45%Nb, and Ti50%Zr on the Pourbaix diagram of titanium at 37 °C (99 °F) indicates that at pH = 6.8, CP titanium and the alloys were well within the thermodynamic stability regime of TiO2 at 37 °C (99 °F), as shown in Fig. 7. The stability of the air- formed oxides in 0.1 M NaCl is confirmed by the spontaneous passivity of titanium, niobium, zirconium and the titanium-niobium- and titanium-zirconium-base alloys.

Fig. 5 Open-circuit potentials of commercially pure (CP) titanium in deaerated 0.1 M NaCl solution and deaerated 5 M HCl solution at 37 °C (99 °F). In 0.1 M NaCl, CP titanium did not exhibit the drop in open-circuit potential that is characteristic of surface activation after oxide dissolution. SCE, saturated calomel electrode

Fig. 6 Open-circuit potentials of Ti-45%Nb and Ti-50%Zr (bulk alloys), initially with air-formed oxides, in deaerated 0.1 M NaCl and 5 M HCl solutions at 37 °C (99 °F). Ti-45% Nb and Ti-50%Zr did not exhibit surface reactivation in 0.1 M NaCl solution (pH ~ 6.8). In contrast, both alloys exhibited surface reactivation in 5 M HCl due to chemical dissolution of the oxide. SCE, saturated calomel electrode

Fig. 7 Potential-pH equilibrium diagram for Ti-H2O system of 37 °C (99 °F). The dissolved titanium species are at an activity of 10-6. Lines a and b define the region of water stability. The experimentally recorded open-circuit potentials (OCP) of commercially pure (CP) titanium, Ti-45%Nb, and Ti-50%Zr

in deaerated 0.1 M NaCl and 5 M HCl solutions before and after surface activation at 37 °C (99 °F) are presented. SCE, saturated calomel electrode. NHE, normal hydrogen electrode

References cited in this section 2. C.R. Brooks, Heat Treatment, Structure and Properties of Nonferrous Alloys, American Society for Metals, 1982 3. R. Boyer, G. Welsch, and E.W. Collings, Ed., Materials Properties Handbook: Titanium Alloys, ASM International, 1994 38. D.J. Blackwood and L.M. Peter, Stability and Open Circuit Breakdown of the Passive Oxide Film on Titanium, Electrochim. Acta, Vol 33 (No. 8), 1988, p 1143–1149 40. D. Kolman and J. Scully, On the Repassivation Behavior of High Purity Titanium and Selected Alpha, Beta, Alpha + Beta Titanium Alloys in Aqueous Chloride Solutions, J. Electrochem. Soc., Vol 143 (No. 6), 1996, p 1847–1859 49. E. Kelly, Anodic Dissolution and Passivation of Titanium in Acidic Media, J. Electrochem. Soc., Vol 126 (No. 12), 1979, p 2064 50. N. Thomas and K. Nobe, The Electrochemical Behavior of Titanium, J. Electrochem. Soc., Vol 116 (No. 12), 1969, p 1748–1751


Effects of Alloying Additions on Titanium Passivity The exact manner in which alloying additions improve passivating oxides is yet unclear. Therefore, possible mechanisms by which alloying additions can enhance the resistance to active anodic dissolution and inhibit the chemical dissolution of passive films are reviewed. The following theories are by no means the only theories with regards to passivity and effects of alloying additions on passivation. Promoting Active/Passive Transition. The alloying addition of small concentrations of palladium to titanium induces spontaneous passivation of titanium-palladium alloys in reducing environments where titanium, by itself, exhibits active/passive transitions, as shown in Fig. 8. (Ref 16, 17, 33). In such cases, spontaneous passivation of titanium is due to enhanced cathodic kinetics associated with hydrogen evolution instead of expanded thermodynamic stability of oxide passivity. The presence of palladium and other noble-metal additions, such as iridium, platinum, and rhodium, at the solution-metal or solution/oxide/metal interface modify the kinetics of hydrogen evolution by: (a) increasing the exchange current density of hydrogen evolution, and/or (b) reducing the cathodic Tafel slope associated with hydrogen evolution. Increased cathodic exchange current densities and/or reduced cathodic Tafel slopes can result in passivity if the resulting OCP is above the active/passive transition potential of the binary alloy (i.e., TiO2 is not converted to TiOOH). Moreover, changes in the cathodic Tafel slope of titanium- palladium alloys would indicate that the ratedetermining step during hydrogen evolution is no longer a slow discharge reaction step, as described for

titanium (Ref 49). Instead, the rate- determining step during hydrogen evolution on titanium-palladium alloys is more consistent with the chemical hydrogen recombination step (Ref 49).

Fig. 8 Polarization measurement of titanium-palladium alloys in acidic sodium chloride solution (deaerated, sweep rate = 0.2 V/min, NaCl = 250 g/L, pH = 0.5, and boiling). SCE, saturated calomel electrode; CP, commercially pure. Source: Ref 51

References cited in this section 16. P.A. Mausli, S.G. Steinemann, and J.P. Simpson, “Properties of Surface Oxides on Titanium and Some Titanium Alloys,” Proceedings of the Sixth World Conference on Titanium (France), P. LaCombe, R. Tricot, and G. Bèranger, Ed., Sociètè Francaise de Métallurgie, 1988 17. N. Khalil and J.S.L. Leach, The Anodic Oxidation of Valve Metals, Part I: Determination of Ionic Transport Numbers by Alpha Spectrometry, Electrochim. Acta, Vol 31 (No. 10), 1986, p 1279–1285 33. I. Dugdale and J. Cotton, The Anodic Polarization of Titanium in Halide Solutions, Corros. Sci., Vol 4, 1964, p 397–411 49. E. Kelly, Anodic Dissolution and Passivation of Titanium in Acidic Media, J. Electrochem. Soc., Vol 126 (No. 12), 1979, p 2064 51. S. Kitayama et al., “Effect of Small Pd Addition on the Corrosion Resistance of Ti and Ti Alloys in Severe Gas and Oil Environments,” Corrosion 92, NACE, 1992


Formation of Amorphous or Vitreous Oxide It has been asserted that alloying additions such as chromium and manganese that reduce passive current densities do so by reducing the density of crystal defects in oxide films and thereby reducing the ionic conductivity across the oxide (Ref 15). For example, the improved corrosion resistance associated with the addition of 12 wt% chromium to iron to form stainless steel can be attributed to the formation of an amorphous iron-chromium-base oxide (Ref 52). The formation of an amorphous oxide or possibly a vitreous oxide film, for example, short- range order and structure, can offer improved resistance to ion migration due to the absence of grain boundaries and dislocations that can act as low-energy conduits for ion migration (Ref 52). Moreover, the formation of either an amorphous or vitreous oxide enhances resistance to corrosion due to the flexibility of the atomic bonds between metallic cations with O2- or OH-. The benefits of bond flexibility are the following: (a) it allows for tolerance of the misfit strain associated with the presence of halide, specifically Cl-, at the oxidemetal interface, and (b) almost all the surface atoms can bond with oxygen or OH- without requiring an almost perfect epitaxial relationship between metal and oxide (Ref 52). Chromium and molybdenum also have been shown to improve the corrosion resistance of base metals such as iron and titanium by inhibiting anodic dissolution and enhancing passivation in reducing environments such as 6 and 12 M HCl, respectively (Ref 9, 53). Researchers associated the enhanced corrosion resistance and the improved passivity due to the formation of a passivating film consisting of double oxyhydroxides of molybdenum-titanium and chromium-titanium, which have a ratio of O2- to OH- of approximately 2. The double oxyhydroxides on molybdenum-titanium and chromium-titanium have been shown to be amorphous in structure, where molybdenum or chromium ions are bonded directly to titanium ions (Ref 9, 53). The amorphous and homogenous structure of the oxyhydroxides may account for the enhanced resistance to anodic dissolution in comparison to the respective molybdenum, chromium, and titanium oxides. The formation of oxyhydroxides will be helpful for construction of homogenous surface films without defects, which act as weak points for corrosion (Ref 53). Furthermore, it has been shown that zirconium, when alloyed with molybdenum and anodically polarized in 12 M HCl, will form complex double oxyhydroxides, where zirconium and molybdenum are covalently bonded as first nearest neighbors (Ref 54). Therefore, the alloying additions of niobium and zirconium may lead to the formation of amorphous oxyhydroxides when alloyed with titanium.

References cited in this section 9. P.Y. Park et al., The Corrosion Behavior of Sputter-Deposited Mo-Ti Alloys in Concentrated Hydrochloric Acid, Corros. Sci., Vol 38 (No. 10), 1996 p 1649–1667 15. D. Tomashov et al., The Passivation of Alloys on Titanium Bases, Electrochim. Acta, Vol 19, 1974, p 159–172 52. A.G. Revesz and J. Kruger, The Role of Noncrystalline Films in Passivation and Breakdown of Passivation, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., The Electrochemical Society, 1978, p 137–155 53. X. Li et al., Spontaneously Passivated Films on Sputter-Deposited Cr-Ti Alloys in 6 M HCl Solution, Corros. Sci., Vol 39 (No. 5), 1997, p 935–948

54. P.Y. Park et al., The Corrosion Behavior of Sputter-Deposited Amorphous Mo-Zr Alloys in 12 M HCl, Corros. Sci., Vol 37 (No. 2), 1995, p 307–320


Reduction of the Potential Gradient across Surface Film Anodic dissolution of metal is controlled by the applied electrode potential difference between metal and electrolyte. On a bare metal surface in contact with solution, the potential difference is located across a thin electrical double layer, the Helmholtz layer, which separates the bare metal surface from the hydrated ions and is approximately 0.5 nm in thickness (Ref 55). However, the presence of a compact oxide surface film results in a multilayered interface consisting of metal/compact surface film/double layer/solution, as shown in Fig. 9. Therefore, on an oxide-covered surface, metal dissolution reaction transpires as a series of consecutive reaction steps: (a) metal ions are transferred from the underlying surface of the metal to the metal- oxide interface, (b) ions migrate through the oxide film, and then, (c) metal ions are transferred across the Helmholtz double layer to the film- solution interface under the applied potential field. It is generally accepted that the last step, the transfer of ions across the Helmholtz double layer to the film-solution interface, which results in the hydration of the metal ions, is the rate- controlling step in metal dissolution across an oxide-covered interface (Ref 55). The potential difference (φc) across the Helmholtz double layer between a thin compact oxide and electrolyte is a function of the overall potential difference between the metal and solution (φm) as well as the geometry and physical characteristics of the multilayered film (Ref 55). The relationship between φc and φm has been defined as the following: (Eq 4) where L is the semiconductor film thickness, εH is the dielectric constant of the Helmholtz layer, εm is the dielectric constant of the metal oxide, and δdl is the thickness of the Helmholtz layer. Therefore, if there is no oxide (surface) film, then φc = φm. However, the presence of a thin oxide film can reduce φc and therefore reduce the driving force necessary for metal-ion transfer across the Helmholtz layer and consequently reduce metallic dissolution. Two factors can be controlled with respect to the oxide: first, oxide thickness (L) and second, the oxide dielectric constant (εm). Thus, it is possible to reduce the potential drop across the Helmholtz layer and, accordingly, the potential field-driven anodic dissolution of the underlying metal by increasing the oxide thickness and/or decreasing the oxide dielectric constant. Consequently, any alloying additions that can increase the compact oxide thickness and/or reduce the dielectric constant of the oxide film can reduce anodic dissolution and enhance corrosion resistance.

Fig. 9 Schematic of the active metal/passive oxide/Helmholtz double layer/solution interfaces that are present on a passivated metal surface

Reference cited in this section 55. N. Sato and G. Okamoto, Electrochemical Passivation of Metals, Eletrochemical Materials Science, J.O.M. Bockris, Ed., Plenum Press, 1981, p 193–245


Enhancement of Atomic Bond Strength The resistance of metals to dissolution and the propensity for passivation are functions of the atomic bond strengths (Ref 56). Metallic-to-metallic atom bonds and bonds between metallic atoms and oxygen or hydroxyl ions form an adsorbed hydroxyl layer (Ref 56). Relative metallic-to-metallic atom bond strength (εM-M) is shown in Eq 5. Similarly, the relative bond strengths between metallic atoms and oxygen can be approximated by the enthalpies (ΔH) of oxygen adsorption at a surface coverage (θ) of zero: (Eq 5) where Z is the coordination number of atom M. The specific enthalpies of oxygen adsorption are provided in Table 3. In comparing the enthalpy of sublimation (ΔHsub) and the enthalpy of oxygen adsorption, one can determine the relative abilities of elements to resist dissolution by creating strong metal-to-metal bonds and to enhance passivity by increasing the bond strengths between metal and oxygen. A high probability of passive oxide formation is indicated by: (a) a high negative energy change associated with oxygen adsorption, for example, metal bonding with oxygen; and (b) the ease by which initial metal-metal bonds can be broken, for

example, low enthalpy of sublimation. A net negative energy change as a result of a metal-metal bond breaking and a metal-oxygen bond forming would indicate the thermodynamic propensity for the formation of oxides. However, the presence of weak metal-metal bonds implies a high rate of metallic dissolution when bare metal is exposed to solution. In contrast, a metal with high metal-metal bond strengths is less likely to undergo dissolution (Ref 56, 57). Such metal will resist the formation of strong bonds with oxygen in favor of forming strong bonds between metallic atoms (Ref 56). Therefore, the mechanisms for stable passivity and stable resistance to dissolution are competitive. As such, two types of alloying elements exist: (a) passivity promoters that possess high enthalpies of oxygen adsorption and relatively low metal-metal bond strength, and (b) dissolution blockers that have high metal-metal bond strengths to inhibit metallic bond failures. However, by adding alloying elements with both characteristics, it would be possible to create an alloy that exhibits both resistance to anodic dissolution and stable passivity. Table 3 Data for enthalpy of adsorption (oxygen), ΔHads, and atomic bond strength, εM-M, for a selection of metals εM-M was calculated from the heat of sublimation of the metals at 25 °C (77 °F). Atomic Element Structure(a) Coordination Heat of adsorption of Heat of sublimation εM-M, 25 °C 77 °F number oxygen (at θ → 0)(b), number (ΔH , or ΔH ), kJ/mole kJ/mole kJ/mole Al fcc 12 326.4 54.4 900 13 Ti hcp 12 469.9 78.3 992 22 Cr bcc 8 396.6 99.1 737 24 Fe bcc 8 416.3 104.1 571 26 Ni fcc 12 429.7 71.6 461 28 Cu fcc 12 338.3 56.4 293 29 Zn hcp 12 130.7 21.8 277 30 Nb bcc 8 725.9 181.5 875 41 Mo bcc 8 658.1 164.5 718 42 Ta bcc 8 782.0 195.5 900 73 W bcc 8 849.4 212.3 819 74 Pt fcc 12 565.3 94.2 285 78 Zr hcp 12 453.0 91.0 NA(c) 140 (a) fcc, face-centered cubic; hcp, hexagonal close-packed; bcc, body-centered cubic. (b) θ, surface coverage. (c) NA, not applicable. Source: Ref 56 A classic example of both alloying types being used is chromium and molybdenum in stainless steels. Chromium, with its high enthalpy of oxygen adsorption, is added to iron to enhance oxide formation on stainless steel, as shown in Table 3. It is possible that chromium easily promotes the nucleation of chromium oxides on the surface of stainless steels at lower anodic overpotentials in comparison to the passive potentials of iron, thereby widening the potential range of passivity of iron-chromium-base alloys. In addition, molybdenum has a high metal-metal bond strength that indicates high resistance to reduce metallic dissolution, as shown in Table 3. Niobium exhibits a heat of adsorption of oxygen of 875 kJ/mole, which would classify niobium as a strong oxide former. Therefore, the alloying addition of niobium to titanium should result in improved passivation characteristics in comparison to pure titanium.

References cited in this section 56. P. Marcus, On Some Fundamental Factors in the Effect of Alloying Elements on Passivation of Alloys, Corros. Sci., Vol 36 (No. 12), 1994, p 2155–2158 57. Y. Okazaki et al., Effect of Alloying Elements on Anodic Polarization Properties of Titanium Alloys in Acid Solutions, Mater. Trans., JIM, Vol 35 (No. 1), 1994, p 58–66


Lowering of the pH of Zero Charge The relationship between the resistance of an oxide to anion uptake and the pH of zero charge (pHpzc) has been investigated on aluminum-base alloys (Ref 58). An important parameter governing the surface charge and, consequently, the adsorption characteristic of an oxide is the pHpzc (Ref 58, 59). The pHpzc of an oxide is the pH at which the surface has no net charge. Therefore, at a pH lower than the pHpzc, the surface has a net positive charge, and aggressive anions are electrostatically attracted to the surface and can be adsorbed, which can possibly lead later to oxide failure. In contrast, at a pH higher than the pHpzc, the surface has a net negative charge, and cations are attracted to the surface. Therefore, alloys containing elements that form oxides, which have relatively low pHpzc, should exhibit superior resistance to aggressive anion uptake. The alloying additions of niobium and zirconium, which exhibit pHpzc of 2.8 and 5.5, respectively, to aluminum, with a pHpzc of 9.4, have been shown to enhance the resistance to anion uptake, for example, increase the pitting potential (Ref 58). Consequently, an alloy of niobium (pHpzc ~ 2.8) to titanium (pHpzc ~ 5), when oxidized, should exhibit superior resistance to chloride uptake and possible oxide failure in comparison to titanium. In contrast, it is possible that zirconium should have no measurable benefit to oxide failure resistance due to the similarity of the pHpzc of titanium and zirconium. A limitation of the pHpzc theory is its inability to predict a pHpzc for mixed oxides.

References cited in this section 58. P. Natishan, E. McCafferty, and G.K. Hubler, Surface Charge Considerations in the Pitting of IonImplanted Aluminum, J. Electrochem. Soc., Vol 135 (No. 2), 1988, p 321–327 59. P.M. Natishan, E. McCafferty, and W. O' Grady, A Study of Passive Film Using X-Ray Photoelectron and X-Ray Absorption Spectroscopy, Applications of Surface Analysis Methods to Environmental/Materials Interactions, R. Baer, G.D. Davis, and C.R. Clayton, Ed., Electrochemical Society, 1991, p 421


Repassivation Behavior of Titanium and Titanium-Base Alloys The electrochemical behavior of bare metal surfaces and their repassivation properties are important factors in environmentally assisted cracking (EAC), erosion-corrosion, and tribological applications. With the exception of gold, all metals oxidize reactive in aqueous environments, and almost all are protected from corrosion by the formation of a surface oxide film or an adsorbed film structure. This passivating behavior is clearly seen in valve metals, including titanium, niobium, zirconium, tantalum, and aluminum, that rely on kinetic barriers to

oxidation instead of thermodynamic stability of the metal for corrosion resistance. Therefore, passivating oxide films are crucial in the technological utility of valve metals. However, the failure of such oxides as a result of mechanical abrasion, anion interaction, or chemical dissolution is a dilemma that must be addressed. The formation of a passivating film transpires in parallel with the active dissolution of the bare metal. During the formation of passivating film, the following steps can occur: (1) adsorbed film formation, (2) formation and growth of islands of oxides to cover the surface, (3) oxide film thickening, and (4) minimization of disorder in the oxide film. Opposing such reactions during passive film formation, the following reactions can occur: (a) active dissolution of metal across the passivating film; (b) dissolution in bare areas, which undercuts the islands of oxides; and (c) allocation of anodic charge to other competing oxidation reactions that do not contribute to oxide formation or consolidation of a compact oxide. Therefore, repassivation of the activated metal surface is achieved when the reaction rates for the formation of passivating film are greater than the rates of reactions opposing stabilization and thickening of the film. However, if the active dissolution mechanism dominates the net anodic reaction at such a site, then this site can become a stable location for localized corrosion. Therefore, an understanding of the mechanisms behind repassivation kinetics as well as the competing reactions are important for understanding the resistance of a metal to EAC, erosion-corrosion, localized corrosion, and tribology. Repassivation Kinetics. During the repassivation of titanium, a minimum of four reactions can occur on the newly formed bare metal surface in deaerated reducing acid (Ref 39, 60): Ti → Ti3+ + 3e-

(Eq 6)

Ti + 2H2O → TiO2 + 4H+ + 4e-

(Eq 7)

H+ + e- → H

(Eq 8)

Ti3+ + H2O → TiO2+ + 2H+ + e-

(Eq 9)

The extent to which the anodic charge is consumed by Eq 6 can be addressed by inductively coupled plasma (ICP) solution analysis and gravimetric mass loss measurements. The thickening of titanium oxide, as described in Eq 7, can be measured by methods including electrochemical impedance spectroscopy (EIS). In contrast, nascent hydrogen formation cannot be measured by the experimental methods that are available to this researcher. However, the hydrogen-evolution reaction step is not possible at potentials greater than E(VNHE) = 0.0 - 0.06 pH. Further oxidation of dissolved Ti3+ to Ti4+ in the form of TiO2+ can be measured by a scratched disk/gold ring oxidation test, as described in Ref 60. The reaction step described by Eq 6 is reported to be the dominant mechanism that occurs during the repassivation of titanium in reducing acids, where greater than 90% of the applied anodic charge prior to steady-state passivation is consumed by the dissolution of metallic titanium into Ti3+ (Ref 40, 60). Moreover, the active anodic dissolution of titanium to Ti3+ occurs both on the bare metal surface and through the reforming oxide layer, due to the large potential gradient across the oxide. The following are possible reaction steps for metallic titanium dissolution to Ti3+ in deaerated solution, as described by Eq 6: Ti → Ti2+ + 2e-

(Eq 10)

Ti2+ + H+ → [Ti3+H]ads

(Eq 11)

[Ti3+H]ads → Ti3+ + H+ + e-

(Eq 12) 3+

where the final product is aqueous Ti at low potentials (Ref 7, 50). The alloying additions of species such as molybdenum, niobium, and zirconium to titanium could reduce active anodic dissolution rates of titanium-base alloys in comparison to pure titanium by enhancing the atomic bond strengths between titanium, molybdenum, niobium, and zirconium. Suffice it to say that molybdenum, niobium, and zirconium alloying additions can be beneficial in reducing the active anodic dissolution of titanium (Ref 56, 57). It is reasonable to assume that alloying additions of molybdenum, niobium, and zirconium to titanium increase the activation energy that is necessary to remove the valence electrons from titanium that are bonded to molybdenum, niobium, or zirconium in comparison to titanium bonded to titanium. This increase in activation energy required to oxidize titanium in titanium- molybdenum, titanium-niobium, and titanium- zirconium alloys is attributable to confirmed stronger d-d-level electron interaction between titanium-molybdenum, titanium-niobium, and titanium-zirconium near-neighbors in comparison to titanium-titanium. Therefore, the reaction rates for the

titanium-molybdenum, titanium-niobium, and titanium-zirconium alloys undergoing active anodic dissolution should be reduced in comparison to pure titanium. Effects of Alloying Additions on Repassivation Kinetics of Titanium. Active anodic dissolution consumed approximately 90% of the total anodic charge for titanium, Ti-45%Nb, and Ti-50%Zr during galvanostatic polarization at 1 A/cm2 (6.5 A/in.2) (Ref 61). Similar high anodic-dissolution-charge consumption was recorded by researchers during repassivation tests of titanium, which indicates that the majority of the applied anodic charge density was consumed by active anodic dissolution of titanium, niobium, and zirconium from the bare surface (Ref 15, 40, 60). However, the anodic charge densities associated with active anodic dissolution diminished to approximately 10% of the total applied charge densities for Ti-45%Nb and Ti-50%Zr at lower anodic current densities approaching steady-state passivation densities, albeit at a high potential. In contrast, titanium still exhibited greater anodic dissolution charge densities at low current densities, which clearly indicates that the rapidly formed oxides on the titanium-niobium and titanium-zirconium binary alloys are more protective in comparison to TiO2 formed on CP titanium. Mechanisms by which Alloying Elements Inhibit Active Anodic Dissolution during Repassivation. It is reasonable to assume that similar dissolution inhibition mechanisms, as described in “Effects of Alloying on Active Anodic Corrosion of Titanium” are also applicable in inhibiting active anodic dissolution during repassivation. Researchers have explored the issue of enhanced resistance to metallic dissolution, from an atomic bonding approach (Ref 56, 57). They hypothesized that the formation of strong covalent bonds between the titanium and alloying species such as molybdenum, niobium, and zirconium atoms results in reduced rates of active metallic dissolution. Therefore, it reasonable to assume that the reaction rates for titaniummolybdenum, titanium-niobium, and titanium-zirconium alloys, which govern the active anodic dissolution, should be diminished, even at high- voltage driving force, in comparison to pure titanium. Once a passive film has formed, the potential gradient across the metal/passive film/double layer/solution interface is defined by Eq 4, which is dependent on the oxide thickness, the Helmholtz double-layer thickness, as well as the dielectric constants of the oxide and the double layer. It can be assumed that the dielectric constant of the mixed oxides on titanium-base alloys will remain relatively constant (Ref 61). Moreover, assuming that the Helmholtz double-layer thickness as well as the dielectric constant of the double layer remain constant, the primary variable in determining the potential gradient across the oxide film is the oxide thickness (dox). An evaluation of the oxide-thickening ratios for titanium, Ti-45%Nb, and Ti-50%Zr revealed that both Ti45%Nb and Ti-50%Zr exhibited higher oxide-thickening ratios, 3.1 and 2.2 nm/V, respectively, in comparison to titanium, which exhibited a Δdox/ΔEapp ratio of approximately 1.1 nm/V (Ref 61). Subsequently, there is a reduced potential gradient driving anodic dissolution of the underlying metal on Ti-45%Nb and Ti- 50%Zr in comparison to CP titanium. It is evident that such a primary condition for the high field model, for example, ~100% charge efficiency during oxide formation, is violated during galvanostatic polarization at high current densities, that is, i > 0.0001 A/cm2 (0.00065 A/in.2) (Ref 61). All of the tested samples exhibited significant active anodic dissolution, as measured by gravimetric mass loss measurements and ICP solution analysis. Therefore, the high field approximation for oxide growth cannot be used to describe the repassivation behavior of titanium and its alloys, including Ti-45%Nb and Ti- 50%Zr.

References cited in this section 7. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, 1st ed., NACE, 1974 15. D. Tomashov et al., The Passivation of Alloys on Titanium Bases, Electrochim. Acta, Vol 19, 1974, p 159–172 39. T.R. Beck, Electrochemistry of Freshly- Generated Titanium Surfaces, Part I: Scraped Rotating Disk Experiment, Electrochim. Acta, Vol 18, 1973, p 807 40. D. Kolman and J. Scully, On the Repassivation Behavior of High Purity Titanium and Selected Alpha, Beta, Alpha + Beta Titanium Alloys in Aqueous Chloride Solutions, J. Electrochem. Soc., Vol 143 (No. 6), 1996, p 1847–1859

50. N. Thomas and K. Nobe, The Electrochemical Behavior of Titanium, J. Electrochem. Soc., Vol 116 (No. 12), 1969, p 1748–1751 56. P. Marcus, On Some Fundamental Factors in the Effect of Alloying Elements on Passivation of Alloys, Corros. Sci., Vol 36 (No. 12), 1994, p 2155–2158 57. Y. Okazaki et al., Effect of Alloying Elements on Anodic Polarization Properties of Titanium Alloys in Acid Solutions, Mater. Trans., JIM, Vol 35 (No. 1), 1994, p 58–66 60. T.R. Beck, Reactions and Kinetics of Newly Generated Titanium Surfaces and Relevance to Stress Corrosion Cracking, Corrosion, Vol 30 (No. 11), 1974, p 408–414 61. S.Y. Yu, “Mechanisms for Enhanced Active Dissolution Resistance and Passivity of Ti Alloyed with Nb and Zr,” Department of Material Science and Engineering, University of Virginia, 1998


References 1. M.J. Donachie, Jr., Titanium: A Technical Guide, 2nd ed., ASM International, 2000 2. C.R. Brooks, Heat Treatment, Structure and Properties of Nonferrous Alloys, American Society for Metals, 1982 3. R. Boyer, G. Welsch, and E.W. Collings, Ed., Materials Properties Handbook: Titanium Alloys, ASM International, 1994 4. T.P. Hoar and D.C. Mears, Corrosion-Resistant Alloys in Chloride Solutions: Materials for Surgical Implants, Proc. R. Soc. (London) A, Vol 294 (No. 1438), 1966, p 486–510 5. A. McQuillan, Titanium Science and Technology, Plenum Press, 1973, p 915–922 6. R. Schutz, An Overview of Beta Titanium Alloy Environmental Behavior, Beta Titanium Alloys in the 1990's, D. Eylon, R. Boyer, and D. Koss, Ed., Minerals, Metals and Materials Society, 1993, p 75–91 7. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, 1st ed., NACE, 1974 8. R.W. Schutz and J.S. Grauman, “Compositional Effects on Titanium Alloy Repassivation Potential in Chloride Media,” Second International Conference on Localized Corrosion, (Orlando, FL) NACE, 1987 9. P.Y. Park et al., The Corrosion Behavior of Sputter-Deposited Mo-Ti Alloys in Concentrated Hydrochloric Acid, Corros. Sci., Vol 38 (No. 10), 1996 p 1649–1667

10. V. Andreeva et al., Investigation of the Corrosion Resistance and the Electrochemical and Mechanical Properties of Titanium-Niobium Alloys, Corrosion of Metals and Alloys, N. Tomashov and E. Nirolybev, Ed., Israel Program for Scientific Translation, Jerusalem, 1966, p 34–47 11. P.J. Bania, Beta-Titanium Alloys and Their Role in Titanium Industry, D. Eylon, R. Boyer, and D. Koss, Ed., TMS, 1993, p 4 12. M. Levy and G. Sklover, Anodic Polarization of Titanium and Titanium Alloys in Hydrochloric Acid, J. Electrochem. Soc., Vol 116 (No. 3), 1969, p 323–328 13. M. Itagaki, R. Oltra, and B. Vuillemin, Quantitative Analysis of Iron Dissolution During Repassivation of Freshly Generated Metallic Surfaces, J. Electrochem. Soc., Vol 144 (No. 1), 1997, p 64–72 14. A. Mazhar, F. Heakal, and A. Gad-Allah, Anodic Behavior of Titanium in Aqueous Media, Corros. Sci., Vol 44, 1988, p 705–710 15. D. Tomashov et al., The Passivation of Alloys on Titanium Bases, Electrochim. Acta, Vol 19, 1974, p 159–172 16. P.A. Mausli, S.G. Steinemann, and J.P. Simpson, “Properties of Surface Oxides on Titanium and Some Titanium Alloys,” Proceedings of the Sixth World Conference on Titanium (France), P. LaCombe, R. Tricot, and G. Bèranger, Ed., Sociètè Francaise de Métallurgie, 1988 17. N. Khalil and J.S.L. Leach, The Anodic Oxidation of Valve Metals, Part I: Determination of Ionic Transport Numbers by Alpha Spectrometry, Electrochim. Acta, Vol 31 (No. 10), 1986, p 1279–1285 18. J. Leach and B. Pearson, The Effect of Foreign Ions upon the Electrical Characteristics of Anodic ZrO2 Films, Electrochim. Acta, Vol 29 (No. 9), 1984, p 1271–1282 19. M.J. Chappell and J.S.L. Leach, Passivity and Breakdown of Passivity of Valve Metals, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 1003–1034 20. E.J. Kelly and H.R. Bronstein, Kinetics and Mechanism of the Hydrogen Evolution Reaction of Titanium in Acidic Media, J. Electrochem. Soc., Vol 131 (No. 10), 1984, p 2232–2238 21. N. Khalil, A. Bowen, and J.S.L. Leach, The Anodic Oxidation of Valve Metals, Part II: The Influence of the Anoidizing Conditions on the Transport Processes During the Anodic Oxidation of Zirconium, Electrochim. Acta, Vol 33 (No. 12), 1988, p 1721–1727 22. T. Shibata and Y. Zhu, The Effect of Film Formation Conditions of the Structure and Composition of Anodic Oxide Films on Titanium, Corros. Sci., Vol 37 (No. 2), 1995, p 253–270 23. J.A. Davies et al., The Migration of Metal and Oxygen During Anodic Film Formation, J. Electrochem. Soc., Vol 112 (No. 7), 1965, p 675–680 24. G.V. Samsonov, The Oxide Handbook, 2nd ed., IFI/Plenum, 1982 25. J. Pan, D. Thierry, and C. Leygraf, Electrochemical Impedance Spectroscopy Study of the Passive Oxide Film on Titanium for Implant Applications, Electrochim. Acta, Vol 41 (No. 7/8), 1996, p 1143– 1153 26. D.G. Kolman and J.R. Scully, Electrochemistry and Passivty of a Ti-15Mo-3Nb-3Al Beta-Titanium Alloy in Ambient Temperature Aqueous Chloride Solutions, J. Electrochem. Soc., Vol 140 (No. 10), 1993, p 2771

27. N. Casillas et al., Pitting Corrosion of Titanium, J. Electrochem. Soc., Vol 141 (No. 3), 1994, p 636–642 28. P. McKay and D. Mitton, An Electrochemical Investigation of Localized Corrosion on Titanium in Chloride Environments, Corrosion, Vol 41 (No. 1), 1985, p 52–61 29. G. Burstein and R. Souto, Observations of Localized Instability of Passive Titanium in Chloride Solution, Electrochim. Acta, Vol 40 (No. 12), 1995, p 1881–1888 30. J. Yahalom and A. Poznansky, Surface Energy and the Stability of the Passive Layer, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 328–336 31. T. Shibata and Y. Zhu, The Effect of Film Formation Potential on the Stochastic Processes of Pit Generation on Anodized Titanium, Corros. Sci., Vol 36(No. 1), 1994, p 153–163 32. T. Watanabe, H. Naito, and Y. Nakamura, Crevice Corrosion Behavior of Commercially Pure Titanium in NaCl-HCl Solutions, J. Jpn. Inst. Met., Vol 50 (No. 9), 1986, p 822–827 33. I. Dugdale and J. Cotton, The Anodic Polarization of Titanium in Halide Solutions, Corros. Sci., Vol 4, 1964, p 397–411 34. J.J. Noel et al., Passive Oxide Film on Titanium in Aqueous Chloride Solution Probed by Electrochemistry, XPS, and In- Situ Neutron Reflectometry, Surface Oxide Films, The Electrochemical Society, 1996 35. R.D. Armstrong and R.E. Firman, Impedance of Titanium in the Active-Passive Transition, J. Electroanal. Chem. Interfacial Electrochem., Vol 34, 1972, p 391–397 36. J.R. Galvele, Present State of Understanding of the Breakdown of Passivity and Repassivation, Passivity of Metals, R.P. Frankenthal and J. Kruger, Ed., Electrochemical Society, 1978, p 285–323 37. C.K. Dyer and J.S.L. Leach, Reversible Reactions Within Anodic Oxide Films on Titanium Electrodes, Electrochim. Acta, Vol 23, 1978, p 1387–1394 38. D.J. Blackwood and L.M. Peter, Stability and Open Circuit Breakdown of the Passive Oxide Film on Titanium, Electrochim. Acta, Vol 33 (No. 8), 1988, p 1143–1149 39. T.R. Beck, Electrochemistry of Freshly- Generated Titanium Surfaces, Part I: Scraped Rotating Disk Experiment, Electrochim. Acta, Vol 18, 1973, p 807 40. D. Kolman and J. Scully, On the Repassivation Behavior of High Purity Titanium and Selected Alpha, Beta, Alpha + Beta Titanium Alloys in Aqueous Chloride Solutions, J. Electrochem. Soc., Vol 143 (No. 6), 1996, p 1847–1859 41. T.R. Beck, Localized Corrosion, B.F. Brown, Ed., NACE, 1974, p 644 42. C. Baes and R. Mesmer, The Hydrolysis of Cations, John Wiley & Sons, 1976, p 147–168 43. B.F. Brown, C.T. Fujii, and E.F. Dahlberg, J. Electrochem. Soc., Vol 116 (No. 218), 1969, p 218 44. M. Stern and H. Wissenberg, The Electrochemical Behavior and Passivity of Titanium, J. Electrochem. Soc., Vol 106 (No. 9), 1959, p 755–759 45. P. Mckay, Crevice-Corrosion Kinetics on Titanium and a Ti-Ni-Mo Alloy in Chloride Solutions at Elevated Temperatures, Corrosion Chemistry Within Pits, Crevices, and Cracks, HMSO Books, 1987

46. R.W. Schutz, “Understanding and Preventing Crevice Corrosion,” Corrosion 91, (Cincinnati, OH), NACE, 1991 47. L.A. Yao and F.X. Gan, Microelectrode Monitoring the Crevice Corrosion in Titanium, Corrosion, Vol 47 (No. 6), 1991, p 420–423 48. H. Shizhong and M. Xiaoxiong, Localized Corrosion in Titanium in Marine Environments, Corrosion and Corrosion Control for Offshore and Marine Construction, Pergamon Press, 1988 49. E. Kelly, Anodic Dissolution and Passivation of Titanium in Acidic Media, J. Electrochem. Soc., Vol 126 (No. 12), 1979, p 2064 50. N. Thomas and K. Nobe, The Electrochemical Behavior of Titanium, J. Electrochem. Soc., Vol 116 (No. 12), 1969, p 1748–1751 51. S. Kitayama et al., “Effect of Small Pd Addition on the Corrosion Resistance of Ti and Ti Alloys in Severe Gas and Oil Environments,” Corrosion 92, NACE, 1992 52. A.G. Revesz and J. Kruger, The Role of Noncrystalline Films in Passivation and Breakdown of Passivation, Passivity of Metals, R

K. Ogle, and M. Wolpers, Phosphate Conversion Coatings, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 712–719

Phosphate Conversion Coatings Kevin Ogle, Irsid, Arcelor R&D (Maizières-lès-Metz, France); Michael Wolpers, Henkel KGaA (Düsseldorf, Germany)

Introduction PHOSPHATING is used in the metalworking industry to treat substrates like iron, steel, galvanized steel, aluminum, copper, and magnesium and its alloys. Most automobile bodies are zinc phosphated prior to painting to increase corrosion resistance and paint adhesion. The cold extrusion of steel would not be economically feasible without phosphating as a lubrifying film. Other applications include providing temporary corrosion resistance for unpainted metal and electrical resistance (Ref 1, 2, 3). This article gives an overview of the types, uses, and theory of phosphate coatings and their formation. It also discusses the composition of phosphating baths, phosphate layers, and their analysis, as well as the process hardware necessary to realize these treatments.

References cited in this section 1. W. Rausch, Die Phosphatierung von Metallen (The Phosphating of Metals), Eugen G. Leuze Verlag, 1988; ASM International and Finishing Publications Ltd., 1990 2. M. Bastian, P. Kuhm, and G. Meyer, Metalloberfläche, Vol 53 (No. 1), p 10–15; Vol 53 (No. 2), p 10– 14; Vol 53 (No. 3), 1999, p 10–14

3. P.E. Tegehall, “The Chemistry of Zinc Phosphating of Steel,” Ph.D. thesis, Chalmers University of Technology, Göteborg, Sweden, 1990

K. Ogle, and M. Wolpers, Phosphate Conversion Coatings, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 712–719 Phosphate Conversion Coatings Kevin Ogle, Irsid, Arcelor R&D (Maizières-lès-Metz, France); Michael Wolpers, Henkel KGaA (Düsseldorf, Germany)

Structure and Function of the Phosphate Film Phosphate conversion coatings are composed of insoluble tertiary metal (Me) phosphates, Me 3(PO4)2 · xH2O. In this formula, Me represents divalent metallic cations of one or more metallic elements. The phosphate layer or coating is formed on a metal substrate by exposing the metal to a phosphating bath. A summary of different types of phosphate layers is given in Table 1. A major distinction can be made between crystalline and amorphous phosphate layers. Figure 1 shows the morphological differences between crystalline and amorphous phosphate layers on zinc coated steel substrates. X-ray diffraction (XRD) spectra support the crystalline and amorphous character of the layer respectively. In Fig. 1(a) (on the left side), the phosphate crystals are clearly visible, giving a characteristic XRD pattern. By contrast, in Fig. 1(b) (on the right side), the amorphous layer is characterized by irregular structure in the scanning electron micrograph (SEM) and no pattern in the XRD spectra. Table 1 Characteristics of different types of phosphate coatings and associated processes (general formulas) Free acid is defined as the number of mL of 0.1 M NaOH required to reach the first endpoint during titration of 10 mL bath sample. Total acid is related to the second endpoint. Amorphous phosphate layers Crystalline phosphate layers Characteristic Alkali No-rinseHeavy phosphating Low-zinc phosphating phosphating phosphating Zn cation Mn cation (iron (dry-in-place Standard-zinc phosphating) phosphating) phosphating(a) 0.1–1.0 0.1–4 1–7 5–30 10 Painted surfaces The EIS approach has several advantages for evaluations of conversion coatings. First, it is a small signal technique, and the electrode is never polarized very far away from its steady-state corrosion potential as the measurement is made. Second, a numerical characterization of coating protectiveness is rendered by this method. For this reason, the technique may be better suited than exposure-test methods for discriminating small differences in corrosion protection among coatings. Third, after the samples have been exposed to develop some pitting damage, the measurement is very rapid. Meaningful characterizations of the impedance response can be obtained from spectra collected between 10 kHz and 10 mHz, which is a measurement that can be completed in a matter of minutes. To shorten the measurement time even further, it may be possible to determine the impedance at one or several properly selected frequencies to estimate coating protectiveness. Finally, EIS measurement equipment, which consists primarily of potentiostat, impedance analyzer, computer, and operating software has become increasingly affordable and portable over the past decade. Electrochemical impedance spectroscopy methods are not without drawbacks. Coating evaluations by EIS are based on a very complex analytical technique. Skill, knowledge, and attention to detail are required to conduct the evaluations and interpret the data properly. Additionally, the exposure protocol involves immersion of the

coated surface in an aggressive aqueous environment. CCCs are not intended to provide long-term corrosion protection under such conditions. Therefore, interpretation of results from these evaluations must be used with this fact in mind. Tests for Self-Healing. Self-healing is an important component of CCC corrosion protection, but rigorously testing for this property is challenging. The familiar example of self-healing ability is the absence of corrosion in a scribed coating after exposure to an aggressive environment. A simple cell, termed the “simulated scratch” cell, mimics this situation, but allows an experimenter to check more rigorously for the key elements of the selfhealing process. These elements include release of inhibitor from the coating, interaction of the inhibitor with the simulated scratch surface, and increased corrosion resistance of the simulated scratch (Ref 69). This cell (Fig. 8) consists of two metal surfaces, normally several square centimeters in area opposing one another and separated by several millimeters. One metal surface is conversion coated, while the other is left bare to simulate a defect or “scratch.” The two metal surfaces are separated by a gap of several millimeters, which is filled with an attacking electrolyte like a dilute chloride solution. The cell may be equipped with a counter and reference electrode, and either the coated or defect sides of the cell can be interrogated using electrochemical methods. Simulated scratch cells have been used to show that certain chromium-free conversion coatings containing can also be self-healing (Ref 80).

Fig. 8 (a) Cross section of the simulated scratch cell. (b) Photograph of an artificial scratch cell modified to make in situ electrochemical measurements to test for self-healing by conversion coatings. Source: Ref 70 Paint Adhesion. Typical aircraft coating systems consist of a CCC foundation layer, a primer coating, and a topcoat (Ref 81). In these systems, the primary function of the CCC is to promote initial and long-term adhesion of the organic coating on the metal substrate. Organic coating adhesion is promoted by surface topography, surface cleanliness, and possibly coating acidity (Ref 23, 82). Long-term interfacial adhesion is fostered by the high corrosion resistance provided by the inhibiting effects of Cr(VI) leached from the primer and the CCC (Ref 1).

Mechanical Stability. Freshly formed coatings are gelatinous and can be damaged easily. As the coating dries, it hardens considerably. CCCs dried for at least 24 h are considerably more robust. Coating weight determinations by chemical stripping of the CCC are nearly impossible to carry out by nitric acid immersion after this time. Hardened coatings will withstand some mechanical damage, but are not nearly as abrasion resistant as anodized coatings. Electrical contact resistance is an issue for metallic electrical connectors. Connections are chromate conversion coated to increase corrosion resistance, but electrical continuity across the connection must not be degraded by the presence of the coating. Chromated metal surfaces exhibit low electrical contact resistance as measured with a flat metal electrode applied with a small impinging load (Ref 75). Since dried CCCs are thin, they are easily breached, allowing metal-to-metal contact. In service, mechanical breaches caused by contacting metal surfaces can sometimes be tolerated without risk of corrosion due to the self-healing nature of the coating. In other cases, mechanical damage is too severe and atmospheric corrosion assisted by galvanic interactions or crevice corrosion will increase the contact resistance to such an extent that rework of the electrical contact is required (Ref 83). Low electrical contact resistance also makes conversion-coated surfaces amenable to various types of spot welding and arc welding procedures. The intrinsic resistivity of CCCs is distinct from electrical contact resistance. Intrinsic resistivity of conversion coatings can be quite high (Ref 84, 85). Direct-current (dc) resistances of CCCs measured with a small (4 × 10 -3 cm2) mercury droplet that does not breach the CCC range from 1010 to 1012 Ω (Ref 85). Aging and Heating. CCC properties are very dynamic once removed from the coating bath. The film transforms from a fragile gel to a robust, damage-tolerant coating in a few hours. Aging effects appear to be due to consolidation in the Cr(III) backbone, which can be detected by EXAFS measurements (Ref 86, 87). How this consolidation occurs is still a matter of speculation, but the structural rearrangement appears to be closely linked to coating dehydration (Ref 86). Shrinkage cracking also appears to be a consequence of backbone consolidation. Backbone rearrangement over the time frame of days to months produces profound changes in coating properties. Cathodic inhibition as measured in cathodic polarization testing is lost in tens of hours (Ref 25). In EIS testing, losses in corrosion resistance begin to be detected in times as short as 20 days after coating formation (Fig. 9) (Ref 86). In an extreme case, after 100 days exposure to ambient laboratory air, there was no evidence of corrosion protection in EIS testing of a 1 minute CCC on 2024-T3 exposed chloride solution. Chromate leaching also diminishes with time even though the Cr(VI) content of the coating remains the same. This well-known behavior may also be linked to backbone consolidation, which traps Cr(VI) in the coating (Ref 87, 89). Ultimately, over tens of years the Cr(VI) content of CCCs diminishes. After seven years, aging under ambient indoor exposure the Cr(VI) CCCs on aluminum alloy 3003 (Al-1.0Mn) was measured by XANES to decrease from about 30% to levels undetectable by XANES (Ref 59).

Fig. 9 Corrosion resistance measured as parallel coating resistance, Rc, from electrochemical impedance spectroscopy experiments of conversion-coated 2024-T3 exposed to aerated 0.5 M NaCl determined after aging in air. Source: Ref 88 As is well known, CCCs are not heat tolerant (Ref 90, 91). Exposure to elevated temperature accelerates dehydration, consolidation of the Cr(III) backbone, and losses in properties described previously. Losses in CCC corrosion protection, Cr(VI) leaching, and increases in shrinkage cracking are profound when coatings are subject to temperatures in excess of 60 °C (140 °F). Corrosion resistance in salt-spray exposure testing and electrochemical testing can be completely lost when CCCs are heated for more than 15 min at temperatures in excess of this 60 °C (140 °F) threshold (Ref 89).

References cited in this section 1. T. Biestek and J. Weber, Electrolytic and Chemical Conversion Coatings, Porticullis Press LimitedRedhill, Surrey, UK, 1976 23. J.A. Treverton and M.P. Amor, Trans. Inst. Met. Finish., Vol 60, 1982, p 92 25. A.E. Hughes, R.J. Taylor, and B.W.R. Hinton, Surf. Interface Anal., Vol 25, 1997, p 223 29. W. Zhang and R.G. Buchheit, J. Electrochem. Soc., Vol 149, 2002, p B357 39. “Coating, Chromate on Aluminum,” Specification 9904151, U.S. Dept. of Energy, 1989 40. “Chemical Conversion Coatings on Aluminum and Aluminum Alloys,” MIL-C- 5541E, NAEC, 1990 59. C.S. Jeffcoate, H.S. Isaacs, J.A.J. Aldykiewicz, and M.P. Ryan, J. Electrochem. Soc., Vol 147, 2000, p 540 68. G.O. Ilevbare and J.R. Scully, J. Electrochem. Soc., Vol 148, 2001, p B196 69. J. Zhao, G.S. Frankel, and R.L. McCreery, J. Electrochem. Soc., Vol 145, 1998, p 2258 70. L. Xia, E. Akiyama, G. Frankel, and R. McCreery, J. Electrochem. Soc., Vol 147, 2000, p 2556 71. J. Sinko, Prog. Org. Coatings, Vol 42, 2001, p 267 72. E. Akiyama, A.S. Markworth, and G.S. Frankel, J. Electrochem. Soc., 2001 73. J.H. Osborne, J. Du, S.R. Taylor, D. Bernard, A. Nercissiantz, and G.P. Bierwagen, “Advanced Corrosion Resistant Aircraft Coatings,” Final Report 6/96-9/00, Contract No. F33615-96-C-5078, U.S. Air Force Research Laboratory, 2000 74. F. Altmayer, Plat. Surf. Finish., Vol 72, 1985, p 36 75. “Chemical Conversion Materials for Coating Aluminum and Aluminum Alloys,” Naval Air Systems Command, Dept. of the U.S. Navy, 1970 76. G.D. Cheever, J. Paint Technol., Vol 41, 1969, p 259 77. R.G. Buchheit, M.W. Kendig, M.A. Martinez, M. Cunningham, and H. Jensen, Corrosion, Vol 54, 1998, p 61 78. F. Mansfeld, S. Lin, S. Kim, and H. Shih, Corros. Sci., Vol 27, 1987, p 997

79. F. Mansfeld and H. Shih, J. Electrochem. Soc., Vol 135, 1988, p 906 80. R.G. Buchheit, S.B. Mamidipally, and P. Schmutz, Corrosion, Vol 58, 2002, p 3 81. R.L. Twite and G.P. Bierwagen, Prog. Org. Coatings, Vol 33, 1998, p 91 82. E.D. Wick, Tool and Manufacturing Engineers Handbook, Society of Manufacturing Engineers, 1983 83. A. Wehr, H.C. Hardee, N.R. Auckland, and D. Klavens, Proc. 48th Electronic Components and Technology Conference, Institute of Electrical and Electronics Engineers, 1998, p 14 84. M.A. Heine and M.J. Pryor, J. Electrochem. Soc., Vol 114, 1967, p 1001 85. G.E. Pike, “Electrical Properties of Alodine 1200,” Internal Memorandum, Sandia National Laboratories, 1981 86. V. Laget, H. Isaacs, C.S. Jeffcoate, and R. Buchheit, Proc. Electrochem. Soc., Vol 92- 26, 2000, p 173 87. V. Laget, H.S. Isaacs, C.S. Jeffcoate, and R.G. Buchheit, ATB Metallurgie, Vol 40–41, 2000, p 295 88. V. Laget, C.S. Jeffcoate, H.S. Isaacs, and R.G. Buchheit, J. Electrochem. Soc., Vol 150 (No. 9), 2003 89. F. Pearlstein and M.R. D'Ambrosio, Plating, Vol 55, 1968, p 345 90. A. Gallacio, F. Pearlstein, and M.R. D'Ambrosio, Met. Finish., Vol 50, 1966 91. A.L. Glass, Mater. Protect., Vol 26, 1968

R.G. Buchheit and A.E. Hughes, Chromate and Chromate-Free Conversion Coatings, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 720–735 Chromate and Chromate-Free Conversion Coatings R.G. Buchheit, The Ohio State University; A.E. Hughes, CSIRO

Chromate-Free Conversion Coatings Due to pressures from health and environmental legislation, there has been a move away from chromatecontaining products over a number of years. The prospect of replacing chromate conversion coatings has brought with it considerable investigation of potential alternatives based on a broad range of chemistries. Furthermore, within each chemical category there is the potential for a broad range of formulations, most of which will not result in viable industrial processes due to processing or performance limitations. A review of all these alternatives is complicated by the range and quality of performance data for these processes, since different processes are targeted toward different parts of the metal-finishing industry that have different performance requirements. Some comparative studies have been carried out and are a good source of performance data, but do not include all the processes described herein (Ref 92, 93, 94). Furthermore, developments in chromate alternatives are progressing rapidly, and those tested in comparative reports may not reflect the current performance of processes. The aim of this section is to give an indication of the market sector for which a class of processes is being developed, some idea of the chemistry and processing involved, an indication of process control, and, finally,

structure and method of deposition of the coating. Readers are referred to several reviews that have been recently published (Ref 95, 96, 97). A total market review of the introduction of replacement technologies was not attempted since this is continually changing as more knowledge is gained about the alternatives. Chromate conversion coatings found a range of applications in industry, especially in aluminum finishing. The development of alternatives is complicated by both the performance and processing requirements of different sectors of industry (Table 3). Currently, several chromate alternatives have gained acceptance in specific sectors of the market. These markets can be divided into those that require protection in the unpainted condition and those that require performance under paint. For the latter category, many alternatives demonstrate good performance characteristics. The aerospace industry falls into the former case, and a drop-in replacement still does not exist in this high-performance end of the market, which has very high standards for corrosion resistance of the unpainted conversion-coated surface in the neutral salt-spray test. Table 3 Major classes of chromate alternatives Industry sector Status(a) Sheet stock for canning Mature Automotive Developing Architectural Developing Cerium-base Aerospace Evaluation Marine Developing Cobalt-base Auto Developing Aerospace Evaluation Tin and galvanized product Developing Molybdenum-base Aerospace Evaluation Hydrotalcites Some sheet product Developing Manganese-base Aerospace Evaluation Aerospace Evaluation Boehmite coatings Automotive Developing Silane coatings Ferrous metals Evaluation Conducting polymers Automotive Al/Mg alloys Developing Self-assembled monolayers (a) Mature, in the industry for a number of years. Developing, may be introduced soon. Evaluation, still undergoing trials General features of these alternative processes are described in the sections that follow. Table 3 lists the major types of chromate alternatives in use or under development and the industries that are currently targeted by the manufacturers of these products. The majority of these processes are still under development. Fluorozirconoic and fluorotitanic acid coatings are the most mature of these replacement technologies with products on the market for a number of years. Two final points should be made, however, with respect to chromate. First, it should be noted that chromate is often considered as a single process suitable for all aluminum alloys under all processing conditions. In reality, this is not the case since different formulations are used for different applications. Indeed there is no single, published database comparing the performance of chromate on a range of alloys cast or wrought in a range of tempers. The available performance data place a strong emphasis on sheet 2024-T3 with some 7075-T6 and 6061-T6 data. Second, conversion coating is a multistep process usually involving both cleaning and deoxidizing/desmutting prior to conversion coating. Over many years, the metal-finishing industry has optimized the pretreatment steps for chromate conversion coating, and it is not surprising that a chromate-base deoxidizer is often used since it sets up a surface more amenable to chromate conversion coating than other deoxidizers. Chromate alternatives may have their own requirements for pretreatment, which may not be the same as the current process steps. These two factors should be taken into account when considering the development of replacements for chromate. Coating type Titanium and zirconium fluorocomplexes


92. B. Meyers, S. Lynn, and E. Jang, “Alternatives to Chromate Conversion Coating at McClellan Air Force Base,” MITRE, 1993 93. “Alternatives to Chromium for Metal Finishing,” Final NCMS Report 0273RE95, National Center for Manufacturing Sciences, 1995 94. “Alternatives to Chrome Conversion Coatings on Aluminum Alloys 2024, 6061, 7075, and Ion Vapor Deposited Aluminum on Steel (Potential Alternatives Report),” Web page, http://www.jgapp.com/parhughhc.htm, National Defense Center for Environmental Excellence, 1998 95. M. Kendig and R. Buchheit, Surface Conversions of Aluminum and Ferrous Alloys for Corrosion Resistance, Proc. Corros./ 2000 Res. Top. Symposium, NACE International, 2000, p 1 96. J. Ouyang and W.L. Harpel, “Monitoring of Coating Weight in Dried-in-Place Non- Chrome Polyacrylamide Based Treatments for Metals,” U.S. Patent No. 5,401,333, 28 March 1995 97. H. Terryn, ATB Metall., Vol 38, 1998, p 41

R.G. Buchheit and A.E. Hughes, Chromate and Chromate-Free Conversion Coatings, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 720–735 Chromate and Chromate-Free Conversion Coatings R.G. Buchheit, The Ohio State University; A.E. Hughes, CSIRO

Chromate-Free Processes Titanium and zirconium fluorocomplexes have a well-established market use in the container, coil coating, aluminum appliance, transportation, magnesium, aerospace, and extrusion industries and are offered by a number of suppliers. More than 100 customers have converted to these titanium and zirconium chemistries from chromate. In addition, National Aerospace and Space Administration (NASA) has used these coatings for a few years on the solid rocket boosters of the space shuttle (2219 alloy) with completely chromium-free paint systems. These coatings also find use in aluminum and titanium fighter jet applications. Earlier versions of these coatings were inorganic, but current versions contain polymeric components (Ref 95). Processing is relatively simple, involving up to eight steps, although five is the most common. These usually include (depending on the manufacturer) an alkaline clean, rinse, coat, rinse, and dry, sometimes at elevated temperature. The conversion coating bath is acidic (pH < 4) and operated at room temperature up to 40 °C (104 °F). The conversion coating solutions generally contain an acidic hexafluoro metal complex (H2ZrF6, or H2TiF6) or Group IVa,b of the CAS classification of the periodic table, oxyfluoro compounds, excess fluoride, and commonly, a polymer. The added polymers improve corrosion resistance and overlayer adhesion. Two examples are polyacrylic acid (Ref 98) or phenolic resins (Ref 99). Other inorganic acids and phosphates may be added to the composition; phosphates have been reported to improve the corrosion resistance of the coating (Ref 100). Process control is through titration for free fluoride, pH, and titanium. Coating weight is determined by x-ray fluorescence (XRF) of titanium or in the zirconium/polymer versions of the coating (Ref 96). XRF determination for zirconium can be done, but it is much less accurate and not particularly reliable. A typical dry-in- place version of this system can be applied in as little as 0.5 s for a roll coater on a coil line, and coating weights are in the range of 1 to 1300 mg/ m2 (0.1 to 120 mg/ft2) (Ref 98). These types of processes demonstrate

excellent performance in standard industry adhesion tests (Ref 98), but generally do not have very good corrosion resistance in the unpainted state (Ref 93, 94). Characterization of the conversion coating indicates a thin, multilayer structure, depicted in Fig. 10. In the absence of polymer additives, the coating consists of three distinct layers. Adjacent to the metal an aluminum oxide develops, then a zirconia (ZrO2) layer develops on top of the aluminum oxide, and this is capped by a mixed Zr/O/F layer. In the presence of a polymeric component, the ZrO 2 and mixed Zr/O/F layers are replaced by a mixed Al/O/Zr/F coating containing the polymer. The polymer tends to be more concentrated toward the surface (Ref 101). Coatings produced in this process are very thin, being typically in the range 10 to 25 nm (0.4 to 1.0 μin.) depending on composition, but may be 500 nm (20 μin.) in some applications.

Fig. 10 Titanium/zirconium fluorocomplexes. The total coating thickness is 15 V

(Eq 5)

Dc = 6.5 + Es

(Eq 6)

δb = 2.1 × Es0.83

(Eq 7) -2

where units of Dp, Dc, δb are nm; Nc is m ; and Es is V.

Fig. 5 Changes in cell number, Nc; cell size, Dc; pore diameter, Dp; barrier layer thickness, δb; and porosity, α, with steady anode potential, Es, obtained in H2C2O4 solutions Hence, the porosity, α = N α = 0.8 × Ea-0.59

/4, of the oxide film can be expressed as: (Eq 8)

The α value decreases sharply with increasing Ea (see Fig. 5). Porous-type anodic oxide films basically consist of amorphous Al2O3, containing small amounts of water and anions of the electrolyte used. These anions distribute across the pore wall and barrier layer as shown in Fig. 6. Oxide films formed in chromic acid solutions, however, consist of almost pure Al2O3.

Fig. 6 Distribution of protons and anions in porous type anodic oxide films. A, Al2O3-x(anion)y·zH2O; B, Al2O3-x(anion)y; C, Al2O3

Reference cited in this section 4. K. Ebihara, H. Takahashi, and M. Nagayama, J. Met. Finish. Soc. Jpn., Vol 34, 1983, p 548; Vol 35, 1984, p 205

H. Takahashi, Aluminum Anodizing, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 736–740 Aluminum Anodizing Hideaki Takahashi, Hokkaido University

Anodizing In general, the anodizing process of an aluminum or aluminum alloy specimen consists of pretreatments (degreasing, etching, and polishing), anodizing, coloring, and sealing. Each step is described in this section. Pretreatments. Degreasing is performed by dipping the specimen in organic solvents, alcohol, acetone, kerosene, or trichloroethylene, or by wiping with cloths permeated with solvent. Etching and polishing are important pretreatments for conversion coating, metal deposition, and organic coating as well as for anodizing, and typical etching and polishing processes are listed in Table 2. Both acid etching and alkaline etching are

possible, and the latter requires a posttreatment of “desmutting” in a nitric acid solution. Electrograining— electrochemical roughening of the surface—is important as a pretreatment in the fabrication of presensitized plates in offset print. Table 2 Procedures of etching and polishing Procedure Acid etching

Solution 10 vol% (40 wt%) hydrofluoric acid

Processing conditions Dip, room temperature bath, 2–5 min

10 vol% (72 wt%) nitric acid

Acid electrograining Alkaline etching Electropolishing

Chemical polishing

Oxide removal

80 vol% water 1–2 wt% hydrochloric acid 4–10 wt% sodium hydroxide 22 vol% (60 wt%) perchloric acid

Power: 8–15 Vac 200–500 A/m2 (0.15–0.30 A/in.2); 20–30 °C (70–85 °F), 5–10 min Dip, 40–90 °C (105–195 °F), 1–10 min, desmut posttreatment: dip, 25–50 vol% nitric acid, ambient temperature, 3–5 min Anodic, 25–30 V, 5–15 °C (40–60 °F), 2–5 min

78 vol% acetic acid 86 vol% (85 wt%) phosphoric acid

Dip, 80–90 °C (175–195 °F), 2–5 min

14 vol% (72 wt%) nitric acid film 0.5 M phosphoric acid

Dip, 85–95 °C (185–205 °F), 10–30 min

0.2 M chromic acid Electropolishing generally enables a smoother surface than chemical polishing. In order to dissolve oxide films on the aluminum, a mixture of chromic acid and phosphoric acid is used, and very little of the aluminum substrate is dissolved. Table 3 shows the chemical composition and thickness of surface films formed by electropolishing, chemical polishing, and chromic acid and phosphoric acid solution treatments (Ref 5). A 3 to 5 nm thick oxide film is formed by these treatments. The oxide films formed by electropolishing contain small amounts of chloride ions, which are produced by the decomposition of perchlorate ions during electropolishing. ) ions. Chemical polishing results in the formation of oxide films containing phosphate and nitrite ( Chromic acid/phosphoric acid treatment after electropolishing replaces the oxide films formed by the electropolishing with oxide films containing Cr (III) and phosphate ions. Table 3 Chemical composition and thickness of surface films formed by electropolishing, chemical polishing, and film removal treatment Treatment Thickness, nm Chemical composition (a) 3.6 AlOx(OH)2.91-2xCl0.09 (x - 0.62)H2O Electropolishing (b) 4.7 AlOx(OH)2.45-2x(PO4)0.17(NO2)0.04·(x - 0.73)H2O Chemical polishing (c) 4.2 AlCr(III)0.43Ox(OH)3.3-2x(PO4)0.33·(x - 0.07)H2O H3PO4/H2CrO4 treatment Treatment details: (a) 13 °C (55 °F), 3 min, 30 V. (b) 85 °C (185 °F), 3 min. (c) 90 °C (194 °F), 30 min. Other experimental conditions are described in Table 2. Typical anodizing procedures are listed in Table 4. Anodizing in sulfuric acid solutions is common, and results in transparent anodic oxide films, which are adequate for dyeing. These films may be dyed into almost any color in practice and can produce strong colors. Oxalic acid solution gives yellowish oxide films, and chromic

acid solution results in gray colors. Table 4 suggests that the steady anode potential, Es, during anodizing strongly depends on the kind of acid used, decreasing in the following order: phosphoric acid, chromic acid, oxalic acid and sulfuric acid. Phosphoric acid gives the highest Es, suggesting the formation of anodic oxide films with thick pore walls and low porosity. Sulfuric acid solution, by contrast, gives the lowest Es, leading to thin pore walls and high porosity. Anodizing in sulfuric acid solutions, therefore, tends to cause “powdering” during anodizing, and phosphoric acid solutions tend to cause “burning.” Table 4 Procedures of anodizing Procedure Sulfuric acid

Oxalic acid Chromic acid Phosphoric acid Hard anodizing

Temperature °C °F 15– 60–75 25

Voltage, Vdc 15–20

Current density A/dm2 A/ft2 0.8–3 7.5–28

Appearance and other characteristics Transparent

20– 30 45– 55 20– 30 0–5

70–85

25

1–1.5

9.5–14

Yellowish

115– 130 70–85

40–50

0.3–1

28–9.5

Grayish

40–100

0.5–2

Light blue

32–40

25–60

2–4

4.75– 18.5 18.5– 37

1 wt% oxalic acid 5 wt% sulfuric acid

22– 25

70–75

25–70

2–3.2

18.5– 29.75

Bronze-black

10 wt% sulfosalicylic acid 8–12 wt% sodium hydroxide

10– 20

50–70

30–70

1–4

9.5–37

Resistant in basic solutions

0–40 32– 105

15–30(a)

3–12

28– 112

Soft, flexible, sulfide included

180

>160

1

9.5

α-alumina

Chemical composition 5–25 wt% sulfuric acid 0.1–5 wt% aluminum sulfate 3–5 wt% oxalic acid 10 wt% chromic acid 3–10 wt% phosphoric acid 10–20 wt% sulfuric acid

Transparent, hardness, >600 HV

0.1–10 wt% aluminum sulfate

Kalcolor

Alkaline anodizing

2–3 wt% hydrogen peroxide

Alternating current anodizing Molten salt anodizing

0.1–0.5 wt% sodium phosphate 15–30 % sulfuric acid 66 mol% potassium bisulfate

355

33 mol% sodium bisulfate (a) Alternating current Anodizing in sulfuric acid solutions containing sulfosalicylic acid leads to the formation of anodic oxide films with brown to blackish colors, the Kalcolor method. Anodizing at temperatures as low as 0 °C gives hard anodic oxide films with a hardness of 500 HV. Anodizing in alkaline solutions containing hydrogen peroxide

gives porous anodic oxide films with a high resistance to alkali, and anodizing with alternative current in sulfuric acid solution gives flexible anodic oxide films. Coloring of anodic oxide films is classified into three groups: integral coloring, dyeing and electrolytic coloring. In integral coloring, many kinds of organic acids, oxalic acid, sulfosalicylic acid, sulfophthalic acid, maleic acid, succinic acid, sulfamic acid, and so forth, are mixed with sulfuric acid, and the anodic oxide films formed in the mixed solutions give brown, grayish, and blackish colors in general (Fig. 7a). The advantage of this method is to obtain anodic oxide films that are highly resistant to the deterioration of color, and integral coloring has been applied extensively in architecture, curtain walls, window frames, and roofing materials.

Fig. 7 Coloring of porous type anodic oxide films on aluminum by (a) integral coloring, (b) dyeing, and (c) electrolytic coloring Porous anodic oxide films have a large internal surface and easily absorb dyestuff on the pore wall (Fig. 7b). Many organic dyestuffs, such as, Alizaline Blue, Alizaline Red-S, Naphtol Green- B, Chromolan Blue-NGG, and Anthraquinone, are used to give flashy colors: gold, yellow, blue, green, and so forth. Some inorganic dyestuffs are also used, and the process involves the precipitation of metal salts with low-solubility products, Ag2Cr2O7, PbS, and Co(OH)3 into the pores. The color range of the inorganic dyeing is more restricted than that of organic dyeing, but the inorganic dyestuff is much more resistant to heat and light. Electrolytic coloring is generally a method where metal is deposited at the bottom of the pores by applying alternative voltage (Fig. 7c). Nickel, cobalt, tin, and copper are mainly deposited to color in bronze and maroon to black. Typical conditions of electrolytic coloring are listed Table 5. The electrolytic coloring is much more advantageous from an energy-consumption viewpoint than the integral coloring and does not require the use of special alloys. The light fastness of the finishes with electrolytic coloring is also excellent. Integral coloring is, therefore, being replaced with electrolytic coloring in architectural applications. Table 5 Procedures of electrolytic coloring Metal Nickel

Composition 30 g/L nickel sulfate

Time, min 10

Voltage, V 20 (ac)

Appearance and other characteristics Blue

1–5

10–14 (dc)

Light brown-black

10

15 (ac)

Black

30 g/L boric acid 15 g/L ammonium sulfate 50 g/L nickel sulfate 25 g/L ammonium chloride

Cobalt

25 g/L boric acid 20 g/L cobalt sulfate 25 g/L boric acid

Copper

15 g/L ammonium sulfate 20 g/L cupric sulfate

10

13 (ac)

Red

Tin

5–6 g/L sulfuric acid 20 g/L stannous sulfate

5–15

5–8 (ac)

Bronze-black

15–20 g/L sulfuric acid 1–2 g/L phenol sulfonic acid Sealing. Porous anodic oxide films without sealing are not protective against the corrosion of the substrate since aggressive ions easily penetrate through pores. The dyestuff also tends to move out, leading to fading. Sealing is essential as a posttreatment to ensure resistance and durability of finishes. Sealing is usually performed by dipping in boiling water. During sealing, anhydrous anodic oxide films react with hot water to convert to a crystalline hydroxide, a pseudoboehmite, as described by: Al2O3 + xH2O = Al2O3 · xH2O

(Eq 9)

where x is 1.5 to 2. The volume expansion by the formation of pseudoboehmite causes the pores to be sealed, as shown in Fig. 8 (Ref 6). A 10 min sealing treatment is sufficient to seal the pores completely in boiling water, and afterward water penetrates across the hydroxide slowly to allow the further progress of the hydration. In addition, the outermost part of anodic oxide films becomes highly crystallized. Pressurized steam enables quick sealing, but the equipment needed is more difficult to operate.

Fig. 8 Process of pore sealing with hydroxide during dipping in boiling pure water A cold sealing, which is performed in solutions containing Ni2+ and F- ions at ambient temperatures, has become popular because of low energy consumption (Ref 7). The reaction during the cold sealing is considered to be: Al2O3 + xNi2+ + yF- + 3H2O = Al2Nix Fy(OH)z + (6 - z)OH-

(Eq 10)

Aluminum-nickel-fluoro-complexes are deposited in the pores to seal them.

References cited in this section 5. H. Takahashi and M. Nagayama, J. Met. Finish. Soc. Jpn., Vol 36, 1985, p 96 6. M. Koda, H. Takahashi, and M. Nagayama, J. Met. Finish. Soc. Jpn., Vol 33, 1982, p 242 7. E.P. Short and A. Morita, Plat. Surf. Finish., Vol 72, 1988, p 102


Testing of Film Properties The morphology of anodic oxide films can be observed by transmission electron microscopy (TEM) and scanning electron microscopy (SEM). In TEM, oxide films are removed from the metal substrate by immersing anodized specimens in a saturated HgCl2 solution to examine Np and Dp of porous anodic oxide films. Thin slices of vertical sections of oxide films are obtained by the ultra thin sectioning technique (UTST) (Ref 8) with a diamond knife ultramicrotome to examine film thickness, crystal structure, and chemical compounds deposited in the pores (Ref 8). An example obtained by TEM with UTST is given in Fig. 9, showing the vertical section of a porous type oxide film formed on aluminum. Focused ion beam (FIB) technique is also used to make slices of oxide films, and confocal scanning laser microscopy (CSLM) is useful in examining the distribution of imperfections in the anodic oxide film as well as determining film thickness (Ref 9).

Fig. 9 TEM image of the vertical section of a porous type anodic oxide film on aluminum. The aluminum specimen was anodized for 5 min at 30 V in 0.16 M oxalic acid solution at 40 °C (140 °F) and then immersed for 1 h on open circuit in the same solution. The hardness of oxide films can be found using microvickers measurements, but the hardness values can be affected by the hardness of the metal substrate, especially for thin oxide films. Nano-indentation methods have become popular in examining the hardness of anodic oxide films accurately and are useful in measuring the hardness at a local small area. Abrasion resistance measurements have been performed using conventional abrasive test instruments, jet assembly, and Taber abraser. It is difficult to compare the abrasion resistance obtained by one method with that by another method quantitatively, but a comparison of data obtained by one method for different specimens gives a significant order of abrasion resistance. Scratching by an atomic force microscope probe is also a method of examining the wear durability at the local micro area of anodic oxide films (Ref 10). Corrosion resistance of anodized aluminum is evaluated by conventional corrosion tests, such as CASS tests, salt spray tests, and other exposure tests. Galvanic corrosion tests are important because the corrosion rate of aluminum in contact with other metallic materials can be much higher than that of aluminum alone in many atmospheres.


8. H. Takahashi, M. Nagayama, H. Akahori, and A. Kitahara, J. Electron Microsc., Vol 22, p. 149–157, 1973 9. S.-M. Moon, M. Sakairi, and H. Takahashi, Proc. Intern. Symp. on Corrosion and Corrosion Protection, Electrochemical Society, 2002, p 91 10. Z. Kato, M. Sakairi, and H. Takahashi, J. Electrochem. Soc., Vol 148, p C790, 2001


References 1. S. Wernick, R. Pinner, and P.G. Sheasby, Ed., The Surface Treatment and Finishing of Aluminum and Its Alloys, 5th ed., Vol I and II, Finishing Publications Ltd, Teddington, Middlesex, England, 1987 2. G.C. Wood, Porous Anodic Films on Aluminum, Oxide and Oxide Films, Vol 2, J.W. Diggle and A.K. Vijh, Ed., Marcel Dekker, 1972 3. S. Tajima, Anodic Oxidation of Aluminum, Advances in Corrosion Science and Technology, M.G. Fontana and R.W. Staehle, Ed., Plenum Press, 1970 4. K. Ebihara, H. Takahashi, and M. Nagayama, J. Met. Finish. Soc. Jpn., Vol 34, 1983, p 548; Vol 35, 1984, p 205 5. H. Takahashi and M. Nagayama, J. Met. Finish. Soc. Jpn., Vol 36, 1985, p 96 6. M. Koda, H. Takahashi, and M. Nagayama, J. Met. Finish. Soc. Jpn., Vol 33, 1982, p 242 7. E.P. Short and A. Morita, Plat. Surf. Finish., Vol 72, 1988, p 102 8. H. Takahashi, M. Nagayama, H. Akahori, and A. Kitahara, J. Electron Microsc., Vol 22, p. 149–157, 1973 9. S.-M. Moon, M. Sakairi, and H. Takahashi, Proc. Intern. Symp. on Corrosion and Corrosion Protection, Electrochemical Society, 2002, p 91 10. Z. Kato, M. Sakairi, and H. Takahashi, J. Electrochem. Soc., Vol 148, p C790, 2001



J.W. Diggle, T.C. Downie, and D.W. Goulding, Chem. Rev., Vol 69, 1969, p 365 S. Kawai, Anodizing and Coloring of Aluminum Alloys, Metal Finishing Information Services Ltd and ASM International, 2002 M.F. Stevenson, Anodizing, Surface Engineering, Vol 5, ASM Handbook, ASM International, 1994, p 482–493 S. Tajima, Anodic Oxidation of Aluminum, Advances in Corrosion Science and Technology, M.G. Fontana and R.W. Staehle, Ed., Plenum Press, 1970 G.C. Wood, Porous Anodic Films on Aluminum, Oxide and Oxide Films, Vol 2, J.W. Diggle and A.K. Vijh, Ed., Marcel Dekker, 1972

R.D. Granata, Surface Modification Using Energy Beams, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 741–749

Surface Modification Using Energy Beams Revised by R.D. Granata, Florida Atlantic University

Introduction SURFACE MODIFICATION, in the context of this article, is the alteration of surface composition or structure by the use of energy or particle beams. Elements may be added to influence the surface characteristics of the substrate by the formation of alloys, metastable alloys or phases, or amorphous layers. Surface-modified layers are distinguished from conversion or coating layers by their greater similarity to metallurgical alloying versus chemically reacted, adhered, or physically bonded layers. However, surface structures are produced that differ significantly from those obtained by conventional metallurgical processes. This latter characteristic further distinguishes surface modification from other conventional processes, such as amalgamation or thermal diffusion. Two different modification methods are discussed. The first, ion implantation, is the introduction of ionized species (usually elements, for example, Ti+) into the substrate using kilovolt to megavolt ion accelerating potentials. The second method, laser processing, is high-power laser melting with or without mixing of materials precoated on the substrate, followed by rapid melt quenching. The advantages of the surface modification approach to promoting corrosion resistance include (Ref 1): • • • •

Alteration of the surface without sacrifice of bulk properties Conservation of scarce, critical, or expensive alloying elements Production of novel surface alloys (unattainable by conventional metallurgical techniques) with superior properties Avoidance of coating adhesion problems

The disadvantages of the surface modification techniques include:

• • • •

Processing materials having deep or hidden contours Cost of processing equipment Substrate sensitivity to high energy input (for example, thermal instability) The thinness (susceptibility to damage) of the coatings

Ion implantation and laser processing are discussed separately in the following sections.

Reference cited in this section 1. E. McCafferty, P.G. Moore, J.D. Ayers, and G.K. Hubler, in Corrosion of Metals Processed by Directed Energy Beams, C.R. Clayton and C.M. Preece, Ed., The Metallurgical Society, 1982, p 1–21

R.D. Granata, Surface Modification Using Energy Beams, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 741–749 Surface Modification Using Energy Beams Revised by R.D. Granata, Florida Atlantic University

Ion Implantation Surface modification by ion implantation is a technique that was derived from the semiconductor industry. Specimens for corrosion research were initially prepared by using high-energy research instrumentation or commercial semiconductor implanters. Equipment for surface modification of metal parts by batch processing has been built, and production capabilities of the equipment have been evaluated for particular applications. Ion implantation is reaching a stage of development in which specific advantages are being recognized and exploited for enhancing the corrosion and wear resistance of critical metal parts (Ref 2). For wear resistance, ion implantation has been used for the hardening and friction reduction of metal surfaces (Ref 3). Related techniques are also used in conjunction with ion implantation to increase the ratio of material introduced into the substrate per unit area, to provide appropriate mixtures of materials, or to overcome other difficulties involved in surface modification by ion implantation alone. These techniques include: •

• • •

Ion beam sputtering: An ion beam of argon or xenon directed at a target sputters material from the target to a substrate; the sputtered material arrives at the substrate with enough energy to promote good adhesion of the coating to substrate. Ion beam mixing: Deposited layers (electroplating, sputtering) tens or hundreds of nanometers thick are mixed and bonded to the substrate by an argon or xenon ion beam. Plasma ion deposition: Ion beams are used to create coatings having special phases, especially ionbeam-formed carbon coatings in the diamond phase or ion-beam-formed boron nitride coatings. Ion beam assisted deposition: Ion beams are combined with physical vapor deposition.

Materials Processing and Product Characteristics. The ion implantation technique consists of introducing the target (substrate material) into the implanter vacuum chamber, vaporizing and ionizing the implant species, and electrostatically accelerating the ionized species into the target. Figure 1 shows schematics of research and semiconductor production ion implantation devices. A production-type implanter for materials surface modification is shown in Fig. 2.

Fig. 1 Schematics of (a) research-type ion implantation system and (b) a production-type semiconductor implanter

Fig. 2 Commercial ion implanter for materials modification. Courtesy of Spire Corporation The implant species may be most any element that can be vaporized and ionized in a vacuum chamber. The material used to generate the implant species is usually in the form of a chloride salt of a metal, a gaseous compound, or the pure element. Solid materials are heated to provide an adequate vapor pressure of the implant species. The vaporized material is ionized and accelerated toward the target. Magnetic fields separate various mass/charge ratios and enable the selection of specific implant species. A selected species impacts the target material and is implanted within, producing a surface-modified substrate. The depth to which the species is implanted depends on the accelerating voltage and the atomic numbers of the implant versus target species (Ref 4). Commonly used acceleration voltages are 20 to 200 keV. The typical average implantation depth (range) is 10 to 100 nm. The distribution of implanted species versus range is roughly Gaussian, with the surface distribution determined by relative sputtering (ion etching) ratios. Figure 3 shows a diagram illustrating calculated concentration profiles of cobalt, aluminum, and boron implanted in iron at 50 keV to a dose of 1017 ion/cm2 (6.45 × 1017 ion/in.2). Implantation causes removal (sputtering) of the surface layers, but the thickness of the sputtered layer is generally less than the implantation thickness. Sputtering rates of elements differ and depend on such factors as substrate composition, chemical bonding, metallurgical structure, beam energy, and angle of the incident beam. In some cases, sputtering rate becomes equal to the implantation rate; an equilibrium is achieved (saturation), and no net addition of ions to the substrate occurs. Maximum concentrations of implanted species may be as high as 50 at.%.

Fig. 3 Concentration profiles for cobalt, aluminum, and boron implants in iron approximated using methods described in Ref 4. Energy: 50 keV. Dose: 1071 ion/cm2 (6.45 × 1071 ion/in.2) The advantages of the ion implantation surface modification process include (Ref 1): • • •

No sacrifice of bulk properties Solid-solubility limit can be exceeded. Alloy preparation is independent of diffusion constants.

• • • • • • •

No coating adhesion problems, because there is no interface No significant change in sample dimensions Depth concentration distribution is controllable. Composition can be changed without affecting grain sizes. Precise location of implanted area(s) Clean vacuum process Allows screening of the effects of changes in alloy composition

The implantation process can be viewed as a means of changing the material properties of an item only at the surface. An item can be fabricated from a material having desirable bulk properties (machinable, low-cost) and can be modified to provide the requisite surface properties (corrosion resistance, wearability). However, the disadvantages of the process must be kept in mind: • • • • •

The ion beam cannot process deep or hidden contours (approximately ±30° incidence). Processing equipment may be too costly. The substrate may be excessively heated by high energy input; for example, the typical energy flux of 2 W/cm2 (13 W/in.2) can heat a substrate 500 °C (930 °F). Inclusions in substrate material may bridge the depth of the surface-modified region, producing sites with undesirable properties. Treatment depth may be inadequate for some applications, resulting in galvanic effects with localized corrosion or in implant layer removal by abrasion or erosion.

Problems associated with thermal effects and sample handling have been discussed and technical solutions proposed (Ref 4, 5, 6). For example, nitrogen-implanted surfaces must be kept below 200 °C (390 °F), the temperature at which outward diffusion of nitrogen occurs. Many sample types have been implanted while keeping their surface temperatures below 100 °C (212 °F) by using heat sinks and by modifying processing conditions. Gaseous Corrosion Properties. Gaseous corrosion or thermal oxidation, in the absence of a liquid phase, is primarily governed by the transport processes in the oxide and the ability of the oxide to resist spalling and breakdown or abrasion and erosion. The ion implantation process has provided a means for studying and modifying the surface layers active in gaseous corrosion. Interesting results have been obtained with implanted materials, as in the following five examples: •

•

• •

•

In one study, the effects of the implanted material were found to persist to a much greater depth than the thickness of the implanted layer (Ref 7). Oxide layer growths of more than 100 times the implant depth showed the persistence of the implanted element effects. An Ni-20Cr alloy developed a complete, protective Cr2O3 scale more rapidly when implanted with cerium, yttrium, calcium, or aluminum, but implanted zirconium or chromium had no effect. Scale adherence was increased by ion implantation with cerium, yttrium, calcium, or zirconium. For yttrium- implanted Ni-20Cr, the weight change was approximately 25% that of unimplanted material after 24 h at 1000 °C (1830 °F) in 101 kPa (1 atm) oxygen. A stainless steel implanted with 1017 ion/cm2 (6.45 × 1017 ion/in.2) of yttrium gave no observable spalling of the oxide after more than 6500 h as compared to unimplanted material, which spalled more than 2 mg/cm2 (Ref 8). Implantation of aluminum into an Fe-Cr- 1.4Al alloy increased the surface aluminum concentration to 10% and promoted the formation of a thick, protective, pure Al2O3 scale (Ref 9). High doses of silicon ion implanted into low- silicon-containing substrates did not provide the desired corrosion resistance associated with their conventional high-silicon counterpart materials. However, ion beam mixing of deposited silicon films with argon at 500 °C (930 °F) produced silicon-rich films on iron or titanium that are thicker than those obtained by ion implantation alone (Ref 10). These thick silicon layers should provide enhanced corrosion resistance. Attempts are being made to produce highly oxidation-resistant MoSi2 by ion beam mixing of molybdenum and silicon layers (Ref 11). Zirconium was implanted with such elements as iron, cobalt, or nickel to help relieve internal stresses within the oxide film that result in the break-away oxidation associated with protective oxide fracture (Ref 12).

Aqueous Corrosion Properties. There are two approaches to using ion implantation for increasing the corrosion resistance of substrates; the first is to enhance passivation characteristics, and the second is to create novel materials. In the first approach, implantation creates a surface composition based on conventional metallurgical experience that reacts to form a passive layer. In the second approach, implantation creates a surface composition and a structure that are difficult, impossible, or impractical to achieve by conventional metallurgy. The two main problems in applying ion implantation methods for creating corrosion-resistant surfaces are producing a surface layer that has a corrosion rate low enough to prevent penetration to the base material during the designed lifetime of the item and producing a surface that is self-repairing. The first problem is approached by introducing elements that enhance corrosion resistance, as in conventional metallurgy—for example, chromium (Ref 13) or molybdenum (Ref 14) in ferrous metals. Another approach is to create combinations of elements in the surface layers that cannot be produced by conventional metallurgical methods—for example, tantalum in iron (Ref 15) or phosphorus in iron-chromium alloys (Ref 16). Tantalum forms intermetallic compounds with iron, but conventional alloys with sufficiently high tantalum content are not readily produced. Other examples are chemically resistant amorphous layers produced, for example, on iron-chromium alloys implanted with phosphorus, carbon, silicon, or boron. These materials passivate spontaneously in hot 25% hydrochloric acid (HCl) (Ref 16). The second problem is approached by the introduction of elements that are not soluble in or removed by the environment, for example, palladium in titanium. Titanium passivation is strongly promoted by palladium and can be alloyed with it to provide lasting self-passivation properties in acidic media. However, palladium is required only at the interface as a catalyst for the hydrogen reduction reaction and is not readily soluble. Thus, it may be more cost-effective to modify the surface by implantation than to use palladium in the bulk alloy. The palladium-assisted passivation of titanium is a persistent effect with research specimens (Ref 17). Substantial efforts have been made in the search for effective applications of ion implantation for enhancing corrosion resistance (Ref 18, 19, 20, 21, 22, 23, 24). The greatest difficulty lies in the thinness of the implanted layer; corrosion-resistant layers must not be entirely consumed or damaged. Many research efforts have attempted to mimic conventional alloy compositions. However, the corrosion rates of conventional alloy systems are usually too high relative to the thinness of the implanted layer. The most successful approaches have involved detailed knowledge of a specific material and environmental corrosion conditions or have created unconventional materials. The palladium-implanted titanium, tantalum-implanted iron, and phosphorus- or boron- implanted iron-chromium systems described previously are good examples. Another example of adapting ion implantation to a specific application is its use for interfacial reactions. Implantation of aluminum or titanium in 1010 steel significantly decreased the cathodic delamination rates of organic coatings from the implanted substrates (Ref 25). Wear resistance enhancement by ion implantation is included in the discussion of surface- modified corrosionresistant surfaces, because erosion, cavitation, and fretting corrosion are important forms of corrosion. Ion implantation has been shown to have a major effect on wear-resistance properties. The combined effects of ion implantation on corrosion and wear resistance are particularly well suited in applications in which bearings or sliding surfaces are subjected to intermittent use or experience long periods of storage. Wear resistance is a property that has shown significant improvement for some substrates after surface modification by ion implantation. Activity in this area has been increasing at research laboratories, and specialized equipment has been developed for commercialization. Some of the work on the development of wear-resistant ion- implanted materials is focusing on quantification and lifetime predictions. Adhesion and abrasion are the most specific wear mechanisms, with corrosive and surface- fracture-related mechanisms making contributions in certain cases. In adhesive wear, surfaces under sufficient local pressure deform plastically and weld at the contact points. When the surfaces separate, the contact point welds break, resulting in the formation of debris and in surface roughening. The oxidized debris causes most of the subsequent wear. Abrasive wear is simply the cutting of one surface by another edged or pointed surface. Abrasive wear decreases as the hardness of the cut surface approaches that of the cutting surface. Abrasion is dramatically lower when both material hardnesses are equal. Adhesive and abrasive wear as well as tribomechanical properties for ion-implanted materials are discussed in Ref 26 and 27, respectively. There are many examples of ion implantation studies for increasing wear resistance of materials. Many studies have used nitrogen-implanted steel because of the readily available high-current nitrogen beams and because direct comparison to thermal nitriding processes is possible. Increases of the order of two to three times in the microhardness of steels due to nitrogen implantation have been reported (Ref 28). However, large reductions in wear rate often result from relatively small increases in microhardness. Extrusion dies for TiO2-pigmented

plastics fabricated from low-carbon steel often wear rapidly, yet nitrogen implantation increases the die lifetime by up to ten times. This modification suggests that the microhardness of the steel has been raised above that of TiO2 (Ref 26). Extended tool life has been demonstrated for plastic molding machines, with estimated ion implantation costs as low as 10% of the ordinary part costs (Ref 3). Nitrogen implantation of a titanium alloy (Ti- 6Al-4V) has been found to decrease the wear rate by more than two orders of magnitude (Ref 29). Data suggest that wear rates have been reduced by almost three orders or magnitude in a titanium alloy designed for hip joint prostheses (Ref 26). Wear reduction can be attributed to factors other than microhardness. The residual carbon films introduced during implantation of other elements may decrease the occurrence of adhesive wear by providing lower friction coefficients. Implantation of titanium or tantalum usually results in the coimplantation or inward diffusion of carbonaceous matter present in the vacuum system. In the case of titanium-implanted steel, the wear reduction has been attributed to the formation of an amorphous iron-titanium-carbon film. Other systems are being studied—examples include tin-implanted iron, yttrium- and nitrogen-implanted chromium steels, and nitrogen implantation in cobalt-cemented carbide. It has been suggested that deep- penetrating interstitially active elements, such as nitrogen, be used to ion beam mix thin coatings of oxide-modifying materials into substrates for synergistic wear reduction (Ref 26). However, another opinion is that implanted ions do not migrate below the surface. Four mechanisms have been identified as contributing to the reduction of sliding wear of ion-implanted metals (Ref 27): • • • •

Production of low-friction surfaces by altering surface chemistry Modification of microstructure to harden the surface Stabilization of microstructure against deleterious work-hardening effects Introduction of residual (compressive) stresses to combat fracture

Impact wear with and without sliding wear was also studied for various materials implanted with nitrogen (Ref 30). The materials used in the study were: • • •

Soft: aluminum alloy 2024, aluminum alloy 7075, and oxygen-free high-conductivity copper (Unified Numbering System, or UNS, C10100) Medium: cartridge brass (UNS C26000) Cu10Sn, aluminum bronze (Cu-18Al), and sintered iron (Fe5Cu-lC) Hard: America Iron and Steel Institute (AISI) 1018 carbon steel and AISI type 304 stainless steel

Impact wear improved significantly, but the combination of impact and sliding wear showed only a very small effect. Interestingly, one of the materials, cartridge brass, showed no significant improvement. Critical Evaluation of Product Utility. Numerous examples have been given to define application areas that may benefit from the use of implantation techniques for materials processing. The advantages and disadvantages relate to the small dimension of implantation depth (nanometers). Significant niches are being created for materials processed by ion implantation and related techniques. These surface modification techniques for industrial applications are perhaps best appraised by considering the information presented in Table 1. Table 1 Successful applications of ion implantation for industrial exploitation Application Material Typical results Significant (400 times) lifetime Orthopedic implants (artificial hip and Ti-6Al-4V increase in laboratory tests knee joints) protection against Bearings (precision bearings for M50 tool steel, 440C Improved stainless steel, or 52100 corrosion, sliding wear, and rollingaircraft) bearing steel contact fatigue Four to six times normal performance Extrusion tools (spinnerettes, nozzles, Various alloys and dies) Three to five times normal life and Punching and stamping tools (pellet D2 tool steel better end products punches for nuclear fuel, scoring dies for cans) Source: Ref 2

Other examples of possible interest include the following. Implantation of nitrogen into as- cast Ti-6Al-7Nb alloy was characterized for in vitro behavior using electrochemical methods and grazing incidence x-ray diffraction (XRD). Results showed favorable repassivation kinetics (Ref 31). Shape memory alloy, NiTi, was implanted with nitrogen and boron to enhance the surface properties, corrosion, wear, and fatigue resistance for dental applications (Ref 32). The corrosion resistance of cobalt-chromium alloy in contact with ultrahighmolecular-weight polyethylene for hip joint replacements was improved using plasma-based ion implantation of nitrogen (Ref 33). The corrosion resistance of uranium in air has been improved by implantation of nitrogen and carbon (Ref 34). The corrosion resistance of uranium and uranium-niobium alloy has been improved by implantation of niobium (Ref 35). Nitrogen plasma source ion implantation improved the corrosion resistance of bearing steel (52100) (Ref 36). The corrosion resistance of 430 stainless steel was improved by combined vapor deposition and ion implantation of silicon and titanium dioxide (Ref 37). Ion implantation and ion beam assisted deposition are considered for improvement of aluminum corrosion (Ref 38).

References cited in this section 1. E. McCafferty, P.G. Moore, J.D. Ayers, and G.K. Hubler, in Corrosion of Metals Processed by Directed Energy Beams, C.R. Clayton and C.M. Preece, Ed., The Metallurgical Society, 1982, p 1–21 2. P. Sioshansi, Thin Solid Films, Vol 118, 1984, p 61–71 3. B.G. Delves, in Ion Implantation into Metals, V. Ashworth, W.A. Grant, and R.P.M. Proctor, Ed., Pergamon Press, 1982, p 126–134 4. G. Dearnaley, J.H. Freeman, R.S. Nelson, and J. Stephen, Ion Implantation, North- Holland, 1973 5. F.A. Smidt and B.D. Sartwell, Nucl. Instrum. Methods Phys. Res. B, Vol 6, 1985, p 70–77 6. A.S. Denholm and A.B. Wittkower, Nucl. Instrum. Methods Phys. Res. B, Vol 6, 1985, p 88–93 7. F.H. Scott, J.S. Punni, G.C. Wood, and G. Dearnaley, Ion Implantation into Metals, V. Ashworth, W.A. Grant, and R.P.M. Proctor, Ed., Pergamon Press, 1982, p 245–254 8. M.J. Bennett, G. Dearnaley, M.R. Houlton, and R.W.M. Hawes, Ion Implantation into Metals, V. Ashworth, W.A. Grant, and R.P.M. Proctor, Ed., Pergamon Press, 1982, p 264–276 9. U. Bernabai, M. Cavallini, G. Bombara, G. Dearnaley, and M.A. Wilkins, Corros. Sci., Vol 20, 1980, p 19 10. A. Galerie and G. Dearnaley, Nucl. Instrum. Methods, Vol 209–210, 1983, p 823–829 11. G. Dearnaley, in Fundamental Aspects of Corrosion Protection by Surface Modification, E. McCafferty, C.R. Clayton, and J. Oudar, Ed., The Electrochemical Society, 1984, p 1–16 12. G.C. Bentini, M. Berti, A. Carnera, G. Della Mea, A.V. Drigo, S. Lo Russo, P. Mazzoldi, and G. Dearnaley, Corros. Sci., Vol 20, 1980, p 27 13. R. Valori, D. Popgoshev, and G.K. Hubler, J. Lubr. Technol. (Trans. ASME), Vol 105, 1985, p 534–541 14. P.D. Hicks and F.P.A. Robinson, Corros. Sci., Vol 24 (No. 10), 1984, p 885–900 15. V. Ashworth, D. Baxter, W.A. Grant, and R.P.M. Proctor, Corros. Sci., Vol 17, 1977, p 947 16. K. Hashimoto, in Amorphous Metallic Alloys, F.E. Luborsky, Ed., Butterworths, 1983, p 471

17. G.K. Wolf, J. Dressler, W. Ensinger, H. Ferber, A. Meger, and P. Munn, in Fundamental Aspects of Corrosion Protection by Surface Modification, E. McCafferty, C.R. Clayton, and J. Oudar, Ed., The Electrochemical Society, 1984, p 42–51 18. C.R. Clayton and C.M. Preece, Ed., Corrosion of Metals Processed by Directed Energy Beams, The Metallurgical Society, 1982 19. E. McCafferty, C.R. Clayton, and J. Oudar, Ed., Fundamental Aspects of Corrosion Protection by Surface Modification, The Electrochemical Society, 1984 20. C.M. Preece and J.K. Hirvonen, Ed., Ion Implantation Metallurgy, The Metallurgical Society, 1980 21. V. Ashworth, W.A. Grant, and R.P.M. Proctor, Ion Implantation into Metals, Pergamon Press, 1982 22. J.K. Hirvonen, Ed., Ion Implantation, Vol 18, Treatise on Materials Science and Technology, Academic Press, 1980 23. F.A. Smidt, Ed., Ion Implantation for Materials Processing, Noyes Data Corporation, 1983 24. G.K. Hubler, O.W. Holland, C.R. Clayton, and C.W. White, Ed., Ion Implantation and Ion Beam Processing of Materials: B, Vol 27, Proceedings of the Materials Research Society, North-Holland, 1984 25. R.D. Granata, M.A. De Crosta, and H. Leidheiser, Jr., in Proceedings of the Ninth National Congress on Metallic Corrosion, Vol 1, National Research Council of Canada, 1984, p 264–268 26. G. Dearnaley, Nucl. Instrum. Methods Phys. Res. B, Vol 7/8, 1985, p 158–165 27. I.L. Singer, in Ion Implantation and Ion Beam Processing of Materials, Vol 27, G.K. Hubler, O.W. Holland, C.R. Clayton, and C.W. White, Ed., Proceedings of the Materials Research Society, NorthHolland, 1984, p 585–595 28. J.B. Pethica, in Ion Implantation into Metals, V. Ashworth, W.A. Grant, and R.P.M. Proctor, Ed., Pergamon Press, 1982, p 147–156 29. R. Hutchings and W.C. Oliver, Wear, Vol 92, 1983, p 143 30. K.K. Shih, Wear, Vol 105, 1985, p 341–347 31. L. Thair, U.K. Mudali, N. Bhuvaneswaran, K.G.M. Nair, R. Asokamani, and B. Raj, Corros. Sci., Vol 44 (No. 11), 2002, p 2439–2457 32. H. Pelletier, D. Muller, P. Mille, and J.J. Grob, Surf. Coat. Technol., Vol 158–159, 2002, p 301–308 33. D. Ikeda, M. Ogawa, Y. Hara, Y. Nishimura, O. Odusanya, K. Azuma, S. Matsuda, M. Yatsuzuka, and A. Murakami, Surf. Coat. Technol., Vol 156, 2002, p 301–305 34. R. Arkush, M.H. Mintz, and N. Shamir, J. Nucl. Mater., Vol 281, 2000, p 182–190 35. W.-H. Luo, R.-F. Ni, and S. Wu, Mater. Prot. (China), Vol 33, 2000, p 6–8 36. J.H. Booske, W. Wang, C. Baum, C. Clothier, N. Zjaba, and L. Zhang, Surf. Coat. Technol., Vol 111, 1999, p 97–102

37. R. Nishimura, K. Yamakawa, J. Ishiga, Y. Matsumoto, and H. Nagano, Mater. Chem. Phys., Vol 54, 1998, p 289–292 38. P.M. Natishan, E. McCafferty, G.K. Hubler, and B.D. Sartwell, Advances in Coatings Technologies for Surface Engineering, Minerals, Metals and Materials Society/AIME, 1997, p 271–285


Laser Surface Processing Lasers with continuous outputs of 0.5 to 10 kW can be used to modify the metallurgical structure of a surface and to tailor the surface properties without adversely affecting the bulk properties. The surface modification can take the following three forms. The first is transformation hardening, in which a surface is heated so that thermal diffusion and solid-state transformations can take place (Ref 39, 40, 41). The second is surface melting, which results in a refinement of the structure due to the rapid quenching from the melt (Ref 40, 41, 42, 43, 44, 45). The third is surface alloying, in which alloying elements are added to the melt pool to change the composition of the surface (Ref 46, 47, 48, 49, 50, 51, 52). The novel structures produced by laser surface melting and alloying can exhibit improved electrochemical behavior.

Materials Processing The processing parameters necessary for laser surface processing depend on the type of modification to be done (for example, transformation hardening or surface alloying) and the properties of the material that are being treated (for example, thermal diffusivity, heat capacity, and transformation temperatures). Figure 4 shows typical ranges of conditions for various processes. The laser power, power density, and interaction time are the primary variables. Other variables, such as the composition of the atmosphere during treatment or the rate of material addition, are determined by the details of the processing—for example, the necessity of shielding against oxidation and the desired thickness, composition, and structure of the surface layer.

Fig. 4 Interaction times and power densities necessary for various laser surface modification processes The laser-beam-modified layer can range in thickness from 0.01 to 5 mm (0.4 to 200 mils), depending on the processing variables, although thicknesses of 0.05 to 1 mm (2 to 40 mils) are more common. The longer the interaction time of the laser beam with the material, the deeper the processed layer will be. Of the processes shown in Fig. 4, the areas labeled “Cladding and surface melting” and “Transformation hardening” delineate process parameters that typically affect the material to depths from 0.5 to 5 mm (20 to 200 mils) and result in metallurgical structures similar to welded structures. The parameters designated “Rapid surface alloying and melting” affect a surface layer only 0.02 to 0.6 mm (0.8 to 24 mils) thick but result in quench rates to 107 K/s and therefore allow for the production of novel metallurgical structures and alloys. The procedure for laser beam surface modification is shown in Fig. 5. A high-power laser beam is brought to a focus near the surface of the workpiece. After a melt pool is established, the laser beam can be rastered across the surface, melting a strip of material with each sweep, until the entire surface has been processed. The processing is performed in an inert gas atmosphere or vacuum to protect the metal from oxidation. In practice, it is more convenient to move the workpiece rather than the laser beam.

Fig. 5 Schematic of the procedure for the laser processing of metals Figure 6 shows a schematic of an apparatus for processing laboratory specimens. The specimen is passed through the laser beam by the rotation of the turntable, and the entire apparatus is translated a fraction of a melt width between passes to process a larger area. The process is varied by changing the power level of the laser beam, the power density (or spot diameter), or the sweep speed. The effects of these variables on the depth of melting for three materials are shown in Fig. 7. As can be seen, the melt depth can be varied by an order of magnitude simply by changing the sweep speed. The width of the melt pass is determined primarily by the diameter of the laser spot and the thermal properties of the material and is very slightly dependent on the sweep speed. The rate at which a surface can be processed is typically approximately 1 cm2/s (0.16 in.2/s) for a variety of conditions and materials.

Fig. 6 Schematic of an apparatus for laser processing

Fig. 7 Effects of laser-processing conditions on melt depth. (a) For three materials at one power level. (b) For steel at two power levels

Product Characteristics Solidification Effects. The heat flow characteristics of such processing can reasonably be described by onedimensional heat flow in which the rate at which energy enters the surface is balanced by the rate at which heat is conducted away by the cool substrate (Ref 53). A number of studies have been devoted to characterizing the temperature profiles that result from the laser and electron beam processing of materials (Ref 39, 41, 54). The character of the temperature profiles experienced by material at various depths below the surface of the metal is shown in Fig. 8. The three reference temperatures indicated represent liquidus (TL), solidus (TS), and solid- state transformation (TR) temperatures. As the laser beam sweeps across the surface of a metal sample, the surface begins to heat. When the melting temperature of the material is exceeded, a melt front is established and moves from the free surface of the sample toward the bulk. Shortly after the beam is turned off (or has moved away), the melt front slows and then stops. It then moves back toward the free surface as a solidification front. The unique characteristics of this solidification front are the high cooling rates (such as 107 K/s) and the steep temperature gradient (such as 105 K/cm). These characteristics promote the formation of metastable compositions and structures.

Fig. 8 View of laser-melted surface and temperature profiles experienced at different points on the surface during laser melting. Liquidus (TL), solidus (TS), and solid-state transformation (TR) temperatures are indicated. Because of this steep temperature gradient, there is only a thin layer of metal that is cool enough to solidify. The material above this layer is above the liquidus temperature, and the material below this layer has already solidified. There is a strong epitaxial character to the resolidification as a result of the steep temperature gradients (Ref 55, 56). This is illustrated in Fig. 9. The two effects that compete during the solidification are the regrowth of grains from the substrate and the heterogeneous nucleation and growth of new grains in this thin layer of chilled metal. For most alloys, grains from the substrate will grow into the chilled zone, and any grains that may have nucleated within the melt pool will either redissolve or be transported by the melt front to the free surface. As an example of this, Fig. 10 shows a β-titanium alloy that has been laser processed.

Fig. 9 Origins of epitaxial resolidification as a result of laser processing. (a) Before laser melting. (b) During resolidification. (c) After quenching

Fig. 10 Planar resolidification of a ß-titanium alloy The epitaxial solidification can have a striking effect on the grain size of the surface layer and on the crystallographic texture. Because some crystal planes grow faster than others—for example, (100) for cubic crystals and (1010) for hexagonal close-packed crystals—those substrate grains having a large component of these orientations in the direction of the temperature gradient will grow at the expense of less favorably oriented grains. As these grains grow, they will tend to squeeze out less favorably oriented grains. This affects the crystallographic texture in two ways. First, there is a higher probability for grains at the free surface to be oriented with the fast-growing direction normal to the free surface, and second, low-angle grain boundaries will occur with a higher frequency than in the substrate. As shown in Fig. 11, when the grain size of the metal before processing is small compared to the depth of melting, there is a coarsening of the grain size, and the texture effects are more pronounced. Examples of these effects are shown in Fig. 9 for a large-grain material and in Fig. 12 for a fine-grain material.

Fig. 11 Effect of epitaxial resolidification on grain size for a material with large original grain size (a and b) and a material with small original grain size (c and d) compared to the melt depth. (a) and (c) Before processing. (b) and (d) After processing

Fig. 12 Coarsening of grain size in a ferritic stainless steel by laser melting During the solidification process, there is generally a redistribution of solute atoms away from the solidification front, so that the interface can remain as close to equilibrium as possible (Ref 57). After an initial transient, this partitioning results in the geometrical destabilization of the solidification front, and dendritic solidification results. For a single-crystal alloy that undergoes rapid surface melting, the dendritic character of the resulting melt is shown in Fig. 13. This effect was documented by the laser melting of single crystals of a nickel-base superalloy (Ref 55). The high-temperature gradient affects the character of the dendritic growth in two ways. First, it restrains crystal growth to the directions of the fast-growing crystal planes, as described previously, and second, it minimizes the amount of segregation because of the short times allowed for lateral diffusion. An example of the dendritic character of laser-processed material is shown in Fig. 14. The diameter of the dendrites can be used to estimate the cooling rates experienced by the material as it resolidified. Higher cooling rates result in less lateral diffusion and smaller dendrites, while lower cooling rates result in more lateral diffusion and larger-diameter dendrites.

Fig. 13 Dendritic solidification of cubic single crystals. The high-temperature gradient resulting from laser melting causes dendrites to grow in the preferred •100• direction

Fig. 14 Dendritic solidification in laser-surface-melted type 304 stainless steel. (a) Surface replica of a polished-and-etched cross section. (b) Scanning electron micrograph of the free surface To the extent that equilibrium cannot be maintained at the solidification front by partitioning, as described in the previous paragraph, more than the equilibrium amount of solute atoms will be incorporated into the solidification structure. An increase in the cooling or solidification rates results in still greater deviations from the equilibrium concentrations (Ref 58, 59). In systems that exhibit solubility at all compositions, for example, the aluminum-manganese system, the solubility limits can be extended. Deviations from equilibrium can also result in the formation of metastable crystal structures when the system has several phases with comparable free energies of formation (Ref 60, 61). Amorphous or glassy surface layers can be produced by laser processing only when the system has several energetically and structurally similar crystalline phases; thus, atoms attaching to the solidification front will have a variety of closely spaced lattice sites with similar energies to choose from. This confuses the long-range order and destroys the crystalline character of the solid. Changing the Chemical Composition. In addition to the structural modifications described in the previous section, the composition of the surface layer can also be changed to form surface alloys whose compositions can be tailored to suit surface requirements. Changes in composition have been accomplished with several techniques (Ref 39, 47, 49, 50, 62). For example, the specimen can be coated before laser processing with a thin layer of the desired alloying element or elements. Such coatings have been applied by vacuum deposition, electroplating, spraying, or the use of a slurry. The alloying elements can also be introduced directly into the melt pool during processing in the form of wire or as an injected powder. When the composition is changed,

care must be taken to ensure that the alloying elements are distributed uniformly in the surface layer. Three mechanisms have been identified for facilitating the mixing of components: liquid-state diffusion, convection currents within the melt pool, and agitation of the melt pool by a laser-sustained plasma (Ref 50). A uniform composition can generally be obtained by processing two or more times, as shown in Fig. 15. The first set of passes disperses the alloying elements to the desired depth and determines the overall composition of the surface alloy. Subsequent passes (using processing conditions selected to obtain slightly shallower melt depths) further homogenize the surface.

Fig. 15 Laser surface alloying of molybdenum into Ti-6Al-4V showing the effectiveness of multiple passes in homogenizing the chemical composition. (a) Single pass. (b) Double pass Effects of Microstructure on Properties. The microstructural changes that result from laser processing can take several forms: a redistribution of major alloy components, a redistribution of any second phases or precipitates, a change in the crystalline character of the surface, or a change in the composition. The melting process results in a leveling of any large-scale composition variations, but the solidification transients can also result in a surface that is enriched in one or more of the alloy components. The high cooling rates during solidification and the accompanying extension of solid-solubility limits often result in either the elimination of second phases or a fine distribution of second phases or precipitates. Alloying may be necessary to take advantage of the higher solubility limits. The steep temperature gradients present during processing generally result in epitaxial resolidification, which introduces some crystallographic texture of the surface grains and grain boundaries (Ref 63). The modification of composition holds the most promise for improving electrochemical behavior, because the surface chemistry can be adjusted so that a solidification structure is obtained on quenching that exhibits more corrosion resistance than the base alloy (Ref 64).

Corrosion Properties The corrosion resistance of laser-processed materials has been studied by several research groups. For example, laser beam processing was used in one study to melt and rapidly solidify type 304 stainless steel (UNS S30400) and ferritic Fe-13Cr-xMo (x = 0, 2, 3.5, or 5) stainless steels (Ref 65). The critical current density for the passivation of laser-processed type 304 stainless steel in 1 N sulfuric acid (H2SO4) decreased two orders of magnitude compared to unprocessed material. The width of the active peaks decreased for ferritic stainless steels versus unprocessed material. Pitting potentials increased by approximately 150 mV for type 304 stainless steel and 5% Mo ferritic stainless steel. Ferritic steels with 0, 2, or 3.5% Mo showed decreases in the pitting potentials. In another case, the corrosion behavior of iron-aluminum bronzes was studied electrochemically in 5% H2SO4 (Ref 66). Substantial beneficial effects were observed and attributed to homogenization of the complex microstructure. Lastly, laser surface alloying was performed to incorporate chromium in Society of Automotive Engineers (SAE) 1018 steel (UNS G10180) and molybdenum in type 304 stainless steel (Ref 67). The SAE 1018-Cr materials passivated similarly to corresponding bulk alloys (5 to 80% Cr). The 304-3Mo material was similar in pitting resistance to type 316 stainless steel (UNS S31600). The 304-9Mo material was superior to type 316 stainless steel and showed no pitting up to oxygen evolution potentials. Table 2 shows some of the results for the 304-Mo materials. Plasma-sprayed titanium coatings on steel were consolidated by laser processing to remove residual porosity, which is the major obstacle to enhanced corrosion resistance (Ref 68). Table 2 Effect of laser surface alloying with molybdenum on pitting potentials of austenitic stainless steels in 0.1 M NaCl

Sample

Composition, % Cr Ni Mo Type 304 18–20 8–10 0 Type 316 16–18 10–14 2–3 18.9 9.1 3.7 304-Mo 11.7 9.6 304-9Mo 19.2 SCE, saturated calomel electrode

Pitting potential (Epit), (V vs SCE) 0.300 0.550 0.500 Did not pit

Critical Evaluation of Product Utility Laser processing provides unique opportunities for producing corrosion-resistant surface layers. Highperformance surface layers can be designed while conserving scarce, expensive, or critical materials. However, the major limitations associated with laser processing are the compatibility of substrates with thermal conduction requirements and the current restriction to planar substrates. Other examples include: • • • •

The intergranular corrosion resistance of Al 6013-T651 was improved by excimer laser surface treatment in nitrogen atmosphere, resulting in formation of AlN (Ref 69). The pitting corrosion resistance of Ti-6Al-4V alloy (UNS R56400/R56401) was improved by excimer laser surface treatment in argon atmosphere and, to a lesser extent, in nitrogen atmosphere (Ref 70). The cavitation erosion resistance of brass was improved by laser surface modification using an alloying powder of NiCrSiB (Ref 71). Laser hardfacing of mild steel AISI 1050 was achieved by flame spraying NiCoCrB powder, followed by laser treatment. Cavitation erosion resistance was increased by factors of 4 and 8 in distilled water and 3.5% NaCl, respectively (Ref 72).

References cited in this section 39. J.W. Hill, M.J. Lee, and I.J. Spalding, Opt. Laser Technol., Vol 6, Dec 1974, p 276 40. E.M. Breinan, B.H. Kear, and C.M. Banas, Phys. Today, Vol 29, 1976, p 11, 44 41. H.E. Cline and T.R. Anthony, J. Appl. Phys., Vol 48, 1977, p 3895 42. T.R. Anthony and H.E. Cline, J. Appl. Phys., Vol 49, 1978, p 1248 43. P.R. Strutt, H. Nowotny, M. Tuli, and B.H. Kear, Mater. Sci. Eng., Vol 36, 1978, p 217 44. G. Staehli and C. Sturzenegger, Scr. Metall., Vol 12, 1978, p 617 45. P.L. Bonora, M. Bassoli, P.L. De Anna, B. Battaglin, G. Della Mea, P. Mazzoldi, and A. Miotello, Electrochim. Acta, Vol 25, 1980, p 1497 46. C.W. Draper, R.E. Woods, and L.S. Meyer, Corrosion, Vol 36, 1980, p 405 47. R.J. Pangborn and D.R. Beaman, J. Appl. Phys., Vol 5, 1980, p 5992 48. P.G. Moore and E. McCafferty, J. Electrochem. Soc., Vol 128, 1981, p 1391 49. D.S. Gnanamuthu, Opt. Eng., Vol 19, 1980, p 783 50. P.G. Moore, in Proceedings of the International Laser Processing Conference, Laser Institute of America, 1981

51. C.W. Draper, Appl. Opt., Vol 20, 1981, p 3093 52. H.W. Bergmann and B.L. Mordike, Z. Werkstofftech., Vol 12, 1981, p 142 53. H.S. Carslaw and J.C. Jaeger, Conduction of Heat in Solids, Clarendon Press, 1959 54. L.E. Grenwald, E.M. Breinan, and B.H. Kear, in Laser-Solid Interactions and Laser Processing—1978, S.D. Ferris, E.J. Leany, and J.M. Poate, Ed., American Institute of Physics, 1979 p 189 55. S.L. Narasimhan, S.M. Copley, E.W. Van Stryland, and M. Bass, ICEE J. Quant. Electr., Vol 13, 1977, p 2D 56. A. Munitz, Metall. Trans. B, Vol 11, 1980, p 563 57. B. Chalmers, Principles of Solidification, Robert E. Krieger, 1977 58. C.W. Draper, S.P. Sharma, J.L. Yeh, and S.L. Bernasek, Surf. Interface Anal., Vol 2 (No. 5), 1980, p 179 59. D.G. Beck, S.M. Copley, and M. Bass, Metall. Trans. A, Vol 13, 1982, p 1879 60. P.T. Sarjeant and R. Roy, Mater. Res. Bull., Vol 3, 1968, p 265 61. J.A. Dantzig and S.H. Davis, Mater. Sci. Eng., Vol 32, 1978, p 199 62. N.B. Dahotre, Ed., Lasers in Surface Engineering, ASM International, 1998 63. L.S. Weinman, J.N. DeVault, and P.G. Moore, in Applications of Lasers in Materials Processing, E.A. Metzbower, Ed., American Society for Metals, 1979, p 259 64. L.E. Grenwald, E.M. Breinan, and B.H. Kear, in Laser-Solid Interactions and Laser Processing—1978, S.D. Ferris, E.J. Leany, and J.M. Poate, Ed., American Institute of Physics, 1979, p 239 65. J.B. Lumsden, D.S. Gnanamuthu, and R.J. Moores, in Fundamental Aspects of Corrosion Protection by Surface Modification, E. McCafferty, C.R. Clayton, and J. Oudar, Ed., The Electrochemical Society, 1984, p 122–129 66. J. Javadpour, C.R. Clayton, and C.W. Draper, in Corrosion of Metals Processed by Directed Energy Beams, C.R. Clayton and C.M. Preece, Ed., The Metallurgical Society, 1982, p 135–145 67. E. McCafferty and P.G. Moore, J. Electrochem. Soc., Vol 133 (No. 6), 1986, p 1090–1096 68. E. McCafferty, G.K. Hubler, P.M. Natishan, P.G. Moore, R.A. Kant, and B.D. Sartwell, Ion Beam or Laser Processing of Metal Surfaces for Improved Corrosion Resistance, Mater. Sci. Eng., 1987 69. C.P. Chan, T.M Yue, and H.C. Man, Mater. Sci. Technol., Vol 18, p 575–580 70. T.M. Yue, J.K. Yu, Z. Mei, and H.C. Man, Mater. Lett., Vol 52, 2002, p 206–212 71. K.F. Tam, F.T. Cheng, and H.C. Man, Proceedings of the Conference on Laser-Solid Interactions for Materials Processing, 2000 MRS Spring Meeting, 25–27 April 2000 (San Francisco, CA), Materials Research Society, 2000, p J3.2.1–J3.2.6

72. C.T. Kwok, F.T. Cheng, and H.C. Man, Proceedings of the Conference on the 20th International Congress on ICALEO 2001: Applications of Lasers and Electro-Optics, 15–18 Oct 2001 (Jacksonville, FL), Laser Institute of America, 2001, p 817–827


Acknowledgment This article has been adapted from R.D. Granata and P.G. Moore, Surface Modification, Corrosion, Vol 13, ASM Handbook (formerly Metals Handbook, 9th ed.), ASM International, 1987, p 498–505.


References 1. E. McCafferty, P.G. Moore, J.D. Ayers, and G.K. Hubler, in Corrosion of Metals Processed by Directed Energy Beams, C.R. Clayton and C.M. Preece, Ed., The Metallurgical Society, 1982, p 1–21 2. P. Sioshansi, Thin Solid Films, Vol 118, 1984, p 61–71 3. B.G. Delves, in Ion Implantation into Metals, V. Ashworth, W.A. Grant, and R.P.M. Proctor, Ed., Pergamon Press, 1982, p 126–134 4. G. Dearnaley, J.H. Freeman, R.S. Nelson, and J. Stephen, Ion Implantation, North- Holland, 1973 5. F.A. Smidt and B.D. Sartwell, Nucl. Instrum. Methods Phys. Res. B, Vol 6, 1985, p 70–77 6. A.S. Denholm and A.B. Wittkower, Nucl. Instrum. Methods Phys. Res. B, Vol 6, 1985, p 88–93 7. F.H. Scott, J.S. Punni, G.C. Wood, and G. Dearnaley, Ion Implantation into Metals, V. Ashworth, W.A. Grant, and R.P.M. Proctor, Ed., Pergamon Press, 1982, p 245–254 8. M.J. Bennett, G. Dearnaley, M.R. Houlton, and R.W.M. Hawes, Ion Implantation into Metals, V. Ashworth, W.A. Grant, and R.P.M. Proctor, Ed., Pergamon Press, 1982, p 264–276 9. U. Bernabai, M. Cavallini, G. Bombara, G. Dearnaley, and M.A. Wilkins, Corros. Sci., Vol 20, 1980, p 19

10. A. Galerie and G. Dearnaley, Nucl. Instrum. Methods, Vol 209–210, 1983, p 823–829 11. G. Dearnaley, in Fundamental Aspects of Corrosion Protection by Surface Modification, E. McCafferty, C.R. Clayton, and J. Oudar, Ed., The Electrochemical Society, 1984, p 1–16 12. G.C. Bentini, M. Berti, A. Carnera, G. Della Mea, A.V. Drigo, S. Lo Russo, P. Mazzoldi, and G. Dearnaley, Corros. Sci., Vol 20, 1980, p 27 13. R. Valori, D. Popgoshev, and G.K. Hubler, J. Lubr. Technol. (Trans. ASME), Vol 105, 1985, p 534–541 14. P.D. Hicks and F.P.A. Robinson, Corros. Sci., Vol 24 (No. 10), 1984, p 885–900 15. V. Ashworth, D. Baxter, W.A. Grant, and R.P.M. Proctor, Corros. Sci., Vol 17, 1977, p 947 16. K. Hashimoto, in Amorphous Metallic Alloys, F.E. Luborsky, Ed., Butterworths, 1983, p 471 17. G.K. Wolf, J. Dressler, W. Ensinger, H. Ferber, A. Meger, and P. Munn, in Fundamental Aspects of Corrosion Protection by Surface Modification, E. McCafferty, C.R. Clayton, and J. Oudar, Ed., The Electrochemical Society, 1984, p 42–51 18. C.R. Clayton and C.M. Preece, Ed., Corrosion of Metals Processed by Directed Energy Beams, The Metallurgical Society, 1982 19. E. McCafferty, C.R. Clayton, and J. Oudar, Ed., Fundamental Aspects of Corrosion Protection by Surface Modification, The Electrochemical Society, 1984 20. C.M. Preece and J.K. Hirvonen, Ed., Ion Implantation Metallurgy, The Metallurgical Society, 1980 21. V. Ashworth, W.A. Grant, and R.P.M. Proctor, Ion Implantation into Metals, Pergamon Press, 1982 22. J.K. Hirvonen, Ed., Ion Implantation, Vol 18, Treatise on Materials Science and Technology, Academic Press, 1980 23. F.A. Smidt, Ed., Ion Implantation for Materials Processing, Noyes Data Corporation, 1983 24. G.K. Hubler, O.W. Holland, C.R. Clayton, and C.W. White, Ed., Ion Implantation and Ion Beam Processing of Materials: B, Vol 27, Proceedings of the Materials Research Society, North-Holland, 1984 25. R.D. Granata, M.A. De Crosta, and H. Leidheiser, Jr., in Proceedings of the Ninth National Congress on Metallic Corrosion, Vol 1, National Research Council of Canada, 1984, p 264–268 26. G. Dearnaley, Nucl. Instrum. Methods Phys. Res. B, Vol 7/8, 1985, p 158–165 27. I.L. Singer, in Ion Implantation and Ion Beam Processing of Materials, Vol 27, G.K. Hubler, O.W. Holland, C.R. Clayton, and C.W. White, Ed., Proceedings of the Materials Research Society, NorthHolland, 1984, p 585–595 28. J.B. Pethica, in Ion Implantation into Metals, V. Ashworth, W.A. Grant, and R.P.M. Proctor, Ed., Pergamon Press, 1982, p 147–156 29. R. Hutchings and W.C. Oliver, Wear, Vol 92, 1983, p 143 30. K.K. Shih, Wear, Vol 105, 1985, p 341–347

31. L. Thair, U.K. Mudali, N. Bhuvaneswaran, K.G.M. Nair, R. Asokamani, and B. Raj, Corros. Sci., Vol 44 (No. 11), 2002, p 2439–2457 32. H. Pelletier, D. Muller, P. Mille, and J.J. Grob, Surf. Coat. Technol., Vol 158–159, 2002, p 301–308 33. D. Ikeda, M. Ogawa, Y. Hara, Y. Nishimura, O. Odusanya, K. Azuma, S. Matsuda, M. Yatsuzuka, and A. Murakami, Surf. Coat. Technol., Vol 156, 2002, p 301–305 34. R. Arkush, M.H. Mintz, and N. Shamir, J. Nucl. Mater., Vol 281, 2000, p 182–190 35. W.-H. Luo, R.-F. Ni, and S. Wu, Mater. Prot. (China), Vol 33, 2000, p 6–8 36. J.H. Booske, W. Wang, C. Baum, C. Clothier, N. Zjaba, and L. Zhang, Surf. Coat. Technol., Vol 111, 1999, p 97–102 37. R. Nishimura, K. Yamakawa, J. Ishiga, Y. Matsumoto, and H. Nagano, Mater. Chem. Phys., Vol 54, 1998, p 289–292 38. P.M. Natishan, E. McCafferty, G.K. Hubler, and B.D. Sartwell, Advances in Coatings Technologies for Surface Engineering, Minerals, Metals and Materials Society/AIME, 1997, p 271–285 39. J.W. Hill, M.J. Lee, and I.J. Spalding, Opt. Laser Technol., Vol 6, Dec 1974, p 276 40. E.M. Breinan, B.H. Kear, and C.M. Banas, Phys. Today, Vol 29, 1976, p 11, 44 41. H.E. Cline and T.R. Anthony, J. Appl. Phys., Vol 48, 1977, p 3895 42. T.R. Anthony and H.E. Cline, J. Appl. Phys., Vol 49, 1978, p 1248 43. P.R. Strutt, H. Nowotny, M. Tuli, and B.H. Kear, Mater. Sci. Eng., Vol 36, 1978, p 217 44. G. Staehli and C. Sturzenegger, Scr. Metall., Vol 12, 1978, p 617 45. P.L. Bonora, M. Bassoli, P.L. De Anna, B. Battaglin, G. Della Mea, P. Mazzoldi, and A. Miotello, Electrochim. Acta, Vol 25, 1980, p 1497 46. C.W. Draper, R.E. Woods, and L.S. Meyer, Corrosion, Vol 36, 1980, p 405 47. R.J. Pangborn and D.R. Beaman, J. Appl. Phys., Vol 5, 1980, p 5992 48. P.G. Moore and E. McCafferty, J. Electrochem. Soc., Vol 128, 1981, p 1391 49. D.S. Gnanamuthu, Opt. Eng., Vol 19, 1980, p 783 50. P.G. Moore, in Proceedings of the International Laser Processing Conference, Laser Institute of America, 1981 51. C.W. Draper, Appl. Opt., Vol 20, 1981, p 3093 52. H.W. Bergmann and B.L. Mordike, Z. Werkstofftech., Vol 12, 1981, p 142 53. H.S. Carslaw and J.C. Jaeger, Conduction of Heat in Solids, Clarendon Press, 1959 54. L.E. Grenwald, E.M. Breinan, and B.H. Kear, in Laser-Solid Interactions and Laser Processing—1978, S.D. Ferris, E.J. Leany, and J.M. Poate, Ed., American Institute of Physics, 1979 p 189

55. S.L. Narasimhan, S.M. Copley, E.W. Van Stryland, and M. Bass, ICEE J. Quant. Electr., Vol 13, 1977, p 2D 56. A. Munitz, Metall. Trans. B, Vol 11, 1980, p 563 57. B. Chalmers, Principles of Solidification, Robert E. Krieger, 1977 58. C.W. Draper, S.P. Sharma, J.L. Yeh, and S.L. Bernasek, Surf. Interface Anal., Vol 2 (No. 5), 1980, p 179 59. D.G. Beck, S.M. Copley, and M. Bass, Metall. Trans. A, Vol 13, 1982, p 1879 60. P.T. Sarjeant and R. Roy, Mater. Res. Bull., Vol 3, 1968, p 265 61. J.A. Dantzig and S.H. Davis, Mater. Sci. Eng., Vol 32, 1978, p 199 62. N.B. Dahotre, Ed., Lasers in Surface Engineering, ASM International, 1998 63. L.S. Weinman, J.N. DeVault, and P.G. Moore, in Applications of Lasers in Materials Processing, E.A. Metzbower, Ed., American Society for Metals, 1979, p 259 64. L.E. Grenwald, E.M. Breinan, and B.H. Kear, in Laser-Solid Interactions and Laser Processing—1978, S.D. Ferris, E.J. Leany, and J.M. Poate, Ed., American Institute of Physics, 1979, p 239 65. J.B. Lumsden, D.S. Gnanamuthu, and R.J. Moores, in Fundamental Aspects of Corrosion Protection by Surface Modification, E. McCafferty, C.R. Clayton, and J. Oudar, Ed., The Electrochemical Society, 1984, p 122–129 66. J. Javadpour, C.R. Clayton, and C.W. Draper, in Corrosion of Metals Processed by Directed Energy Beams, C.R. Clayton and C.M. Preece, Ed., The Metallurgical Society, 1982, p 135–145 67. E. McCafferty and P.G. Moore, J. Electrochem. Soc., Vol 133 (No. 6), 1986, p 1090–1096 68. E. McCafferty, G.K. Hubler, P.M. Natishan, P.G. Moore, R.A. Kant, and B.D. Sartwell, Ion Beam or Laser Processing of Metal Surfaces for Improved Corrosion Resistance, Mater. Sci. Eng., 1987 69. C.P. Chan, T.M Yue, and H.C. Man, Mater. Sci. Technol., Vol 18, p 575–580 70. T.M. Yue, J.K. Yu, Z. Mei, and H.C. Man, Mater. Lett., Vol 52, 2002, p 206–212 71. K.F. Tam, F.T. Cheng, and H.C. Man, Proceedings of the Conference on Laser-Solid Interactions for Materials Processing, 2000 MRS Spring Meeting, 25–27 April 2000 (San Francisco, CA), Materials Research Society, 2000, p J3.2.1–J3.2.6 72. C.T. Kwok, F.T. Cheng, and H.C. Man, Proceedings of the Conference on the 20th International Congress on ICALEO 2001: Applications of Lasers and Electro-Optics, 15–18 Oct 2001 (Jacksonville, FL), Laser Institute of America, 2001, p 817–827

Porcelain Enamels, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 750–754

Porcelain Enamels Introduction PORCELAIN ENAMELS are glass coatings applied primarily to products or parts made of sheet steel, cast iron, or aluminum to improve appearance and to protect the metal surface. Porcelain enamels are distinguished from other ceramic coatings on metallic substrates by their predominantly vitreous nature and the types of applications for which they are used. These coatings are differentiated from paint by their inorganic composition and coating properties. They are fused to the metallic substrate at temperatures above 425 °C (800 °F) during the firing process. The most common applications for porcelain enamels are major appliances, water-heater tanks, sanitary ware, and cookware. In addition, porcelain enamels are used in a wide variety of coating applications, including chemical-processing vessels, agricultural storage tanks, piping, pump components, and barbecue grills. They also are used for coatings on architectural panels, signage, specially executed murals, and substrates for microcircuitry. Porcelain enamels are selected for products or components where there is a need for one or more special service requirements that they can provide. These include chemical resistance, corrosion protection, weather resistance, abrasion resistance, specific mechanical or electrical properties, appearance or color needs, cleanability, heat resistance, or thermal-shock capability.

Porcelain Enamels, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 750–754 Porcelain Enamels

Types of Porcelain Enamels Porcelain enamels for sheet steel and cast iron are classified as either ground-coat or cover-coat enamels. Ground-coat enamels contain metallic oxides, such as cobalt oxide and/or nickel oxide, that promote adherence of the glass/enamel to the metal substrate. Cover-coat enamels are applied over fired ground coats to improve the appearance and properties of the coating. Also, cover-coat enamels may be applied over unfired ground coats, with both coats being fired at the same time (i.e., two-coat/one-fire systems). Cover coats may also be applied directly to properly prepared decarburized steel substrates. The color of ground coats is limited to various shades of blue, black, brown, and gray. Cover coats— which may be clear, semiopaque, or opaque— may be pigmented to take on a great variety of colors. Colors may also be smelted into the basic coating material. Opaque cover coats are usually white. Porcelain enamels for aluminum are normally one-coat systems. They are applied only by spraying. When two coats are desired, the first coat can be any color. Frits for porcelain enamels for aluminum are usually transparent and may be colored and opacified with inorganic pigments to produce the desired appearance.


Frits for Porcelain Enameling The basic material of the porcelain-enamel coating is called frit. Frits are smelted complex borosilicate glasses that are produced by quenching a molten glassy mixture. Frits generally consist of from 5 to 20 components, which are thoroughly mixed together and melted into a glassy system. The molten glass is then quenched to a friable (easily broken up) condition by being either poured into water or rolled into a thin sheet between watercooled rolls. When quenched in water, the frit is dried before use. The sheet is ordinarily shattered into small flakes by mechanical means before use or shipment. Because porcelain enamels are usually designed for specific applications, the compositions of the frits from which they are made vary widely. A number of compositions of frits for enamels for sheet steel, cast iron, and aluminum are discussed in the paragraphs that follow; however, many variations of these compositions are used commercially. Enamel Frits for Sheet Steel. The frit compositions given in Table 1 are classified as alkali borosilicates and are used as ground coats on sheet steel. One of the functions of ground-coat enamels is to promote bond to the steel substrate. The addition of cobalt oxide and/or nickel oxide promotes adherence to the substrate. Compositions differ, depending on the end-use application of the enameled product. For example, acid resistance is obtained by the addition of titanium dioxide and high levels of silicon dioxide with a corresponding decrease in the boron level. Resistance of the enamel to alkalis or to water can be improved by adding zirconium oxide or aluminum oxide (usually as zircon or alumina) to the frit while maintaining a high content of silicon dioxide. It is common practice to blend soft (low-melting temperature), medium, and hard (high melting temperature) frits to obtain the maximum desired properties. Table 1 Melted-oxide compositions of frits for ground-coat enamels for sheet steel Constituent Composition, wt% Regular blue-black Alkali-resistant Acid-resistant Water-resistant enamel enamel enamel enamel 37.17 42.02 56.44 48.00 SiO2 22.21 18.41 14.90 12.82 B2O3 18.44 15.05 16.59 18.48 Na2O 0.99 2.71 0.51 … K2O … 1.06 0.72 1.14 Li2O 9.34 4.47 3.06 2.90 CaO … 2.29 … … ZnO 4.53 4.38 0.27 … Al2O3 … 5.04 … 8.52 ZrO2 … … 3.10 3.46 TiO2 … 0.07 0.39 … CuO 1.58 1.39 1.12 0.52 MnO2 1.37 1.04 … 1.21 NiO 0.65 0.93 1.27 0.81 Co3O4 1.15 0.68 … 0.20 P2O5 2.57 2.75 1.63 1.94 F2 9.24 8.59 … … BaO Weather resistance has been shown to be a function of acid resistance. Porcelain enamels for use outdoors are made from various types of frits that produce the necessary acid resistance along with the specific color desired. Resistance to thermal shock and to high temperature is obtained by controlling the coefficient of thermal

expansion of the glass coating; this is accomplished by adjusting the frit composition and adding refractory materials, such as silicon dioxide, aluminum dioxide, and zirconium oxide, to the mill formula. Cover coats for sheet steel are applied over ground coats or directly to properly prepared decarburized steel. Compositions of frits for cover- coat enamels are shown in Table 2. Electrostatic dry powder cover coats may be applied over an electrostatic dry powder base coat; then the entire two-coat/one-fire system is matured in a single firing. Electrostatic dry powder cover coats may also be applied over fired ground coats in a twocoat/two-fire process. Table 2 Melted-oxide compositions of frits for cover-coat enamels for sheet steel Constituent Composition, wt% Titania white enamel Semiopaque enamel Clear enamel 44.67 44.92 54.26 SiO2 14.28 16.40 12.38 B2O3 8.27 8.67 6.55 Na2O 6.99 8.12 11.32 K2O 0.98 0.45 1.14 Li2O … 0.74 … ZnO 1.98 3.34 1.40 ZrO2 0.31 0.16 … Al2O3 18.49 13.05 10.04 TiO2 1.32 0.88 … P2O5 0.5 … … MgO 2.21 3.27 2.91 F2 Cover-coat enamels fall into three groups: clear, semiclear or semiopaque, and opaque. The clear frits do not contain high levels of titanium dioxide, and they remain clear when fired. The semiopaque frits are translucent and are used for light colors. The opaque frits, which contain high levels of titanium dioxide, become white upon firing. Cover-coat enamels made from titanium-dioxide-opacified frits are generally quite acid resistant; even in amounts too small to provide any opacity, titania imparts acid resistance. For alkali resistance, zirconium oxide is a desirable constituent. Clear frits containing 8 to 11% TiO2 are used for strong to medium-strength colors. Semiopaque frits containing 12 to 15% TiO2 are used for medium-strength colors, and opaque frits containing 17 to 20% TiO2 are used for pastel colors. Enamel Frits for Cast Iron. Compositions of frits for enamels for cast iron vary, depending on whether the frit is applied by the dry process or by the wet process (Table 3). Dry-process enamels are commonly used for large cast-iron plumbing fixtures, such as bathtubs and sinks, where the brilliance of the enamel and its ability to cover small surface irregularities in the casting itself are useful characteristics. Acid resistance is imparted to these enamels by reducing the alumina content, increasing silica, and adding up to about 8% TiO2. Ground coats are normally used to fill surface voids in castings. Ground coats for wet-process enamels often are mixtures of frit, enamel reclaim, and refractory raw material used at very low application weight. Ground coats for the dry-process method are applied and fused to thin, viscous coatings that protect the casting surface from excessive oxidation while it is heated to enameling temperature. Table 3 Melted-oxide compositions of frits for enamels for cast iron Constituent Composition, wt% Ground coats LeadNon-leadbearing bearing 77.7 63.4 SiO2 6.8 10.9 B2O3 4.3 7.1 Na2O … 2.4 K2O 4.0 … PbO

Cover coats Zirconiumopacified(a) 28.0 8.8 10.0 4.1 17.8

Titaniaopacified(a) 51.4 9.6 10.1 3.0 …

Titaniaopacified(b) 44.80 13.14 6.48 7.67 0.94

… 0.6 8.7 1.9 0.55 CaO … … 6.1 0.9 … ZnO 7.2 15.1 4.5 2.1 … Al2O3 … … … … 19.74 Sb2O3 … … 6.1 … 1.64 ZrO2 … … … 18.4 … TiO2 … 0.5 5.9 2.6 5.04 F2 (a) For dry process. (b) For wet process Enamel frits for aluminum are usually based on lead silicate or on cadmium silicate, but they may also be based on phosphate, vanadium, or barium. Table 4 gives the compositions of some frits for aluminum. Enamels for aluminum have a high gloss, good acid and weather resistance, and good mechanical properties. They melt at relatively low temperatures. Table 4 Melted-oxide compositions of frits for enamels for aluminum Constituent Composition, wt% Lead-base enamel PbO 14–45 SiO2 30–40 Na2O 14–20 K2O 7–12 Li2O 2–4 B2O3 1–2 Al2O3 … BaO 2–6 P2O5 2–4 F2 … TiO2 15–20 Sb2O5 …

Non-lead-base enamel … 30–40 20–25 7–11 3–5 1–2 … 3–5 2–4 … 15–20 2–5

Barium enamel … 25 20 25 … 15 3 12 … … … …


Properties of Porcelain Enamels Porcelain enamels possess an array of properties that can be specially designed and formulated for a variety of demanding environments. Some of the special properties are summarized in the paragraphs that follow. The Porcelain Enamel Institute bulletins in the Selected References have more detailed information. Appearance for Indoor Exposure. Where corrosive attack is unlikely and appearance is the principal requirement, enamel selection and processing are directed toward providing reproducible color matching and optimal gloss and smoothness. Often, different enamels are used on different parts of the same product to ensure the best balance of properties and cost, particularly if high volume is involved. For example, somewhat different enamel compositions are used for range tops and dishwasher liners because of differing property and appearance requirements, even though these parts may be processed in the same plant. Appearance standards, in particular, are established to respond to end-use requirements. Small surface defects may be tolerated in areas not heavily used or readily seen in the finished part, provided that they do not affect basic serviceability. Thermal shock intensifies the effect of elevated temperature, as does operation under severe temperature gradients. Enamels are formulated so that expansion characteristics place the enamel in compression under

service conditions. If combinations of mechanical stress and elevated temperatures place the enamel in tension, crazing forms a pattern of fine cracks perpendicular to the tensile stress. The resistance of an enamel to thermal shock varies inversely with its thickness. Thermal-shock failure occurs when the temperature gradient perpendicular to the surface is large enough to cause excessive differential shrinkage and tensile stress. Thermal- shock resistance also depends on the design and section thickness of the coated part. Flexing of the metal due to localized thermal gradients parallel to the surface can produce bending and tensile stresses in the coating; therefore, an increase in the strength or rigidity of the part increases the resistance of the coating to thermal shock. Most porcelain enamels applied at conventional thickness can withstand abrupt temperature changes of 110 to 165 °C (200 to 300 °F). Abrasion Resistance. The hardness of porcelain enamels ranges from 3.5 to 6.0 on the Mohs scale. Porcelain enamels show a high degree of abrasion resistance. Under abrasive test conditions where plate glass retains 50% specular gloss, porcelain-enamel compositions retain from 35 to 85% specular gloss. Subsurface abrasion resistance varies with processing variables that affect the bubble structure of the enamel, that is, gas bubbles frozen in during cooling of the enamel. A decrease in abrasion resistance occurs with an increase in the number or size of gas bubbles. Enamel compositions are available that contain crystalline particles (from mill additions or devitrification heat treatment) that increase abrasion resistance as much as 50%.


Corrosion Resistance of Porcelain Enamels Porcelain enamels possess excellent resistance to corrosion in a variety of environments. Chemical Resistance. Porcelain enamel is widely used because of its resistance to household chemicals and foods. Mild alkaline or acid environments are generally involved in household applications. Table 5 presents examples of corrosive environments for which porcelain enamels are widely used for long periods of service. Special enamel compositions are available to resist most acids, except for hydrofluoric or concentrated phosphorics, to 230 °C (450 °F). These compositions also resist alkali concentration to pH 12 at temperatures to 100 °C (212 °F). Table 5 Applications in which porcelain enamels are used for resistance to corrosive environments Application

Corrosive environment Temperature pH °C °F To 49 To 120 5–9 Bathtubs To 100 To 212 12 Chemical ware To 100 To 212 1–2 175– 350–450 1–2 230 To 71 To 160 11 Home laundry equipment 21–66 70–150 2– Range exteriors 10 Range oven liners, 66–315 150–600 2– 10 conventional 66–590 150– 2– Range burner grates 1100 10 -18 to 0–150 2– Refrigerators 66 10

Corrosive medium Water, cleansers Alkaline solutions All acids except hydrofluoric Concentrated sulfuric acid, nitric acid, and hydrochloric acid Water, detergents, bleach Food acids, cleaners Food acids, cleaners Food acids, cleaners Food acids, cleaners

Kitchen sinks

To 71

To 160

2– 10 5–8

Food acids, water, cleaners

Water To 71 To 160 Water heaters Depends on application(a) Industrial heat exchangers (a) Applications include coal- and oil-fired boilers and black liquor evaporators; corrosive media would include ash from coal-fired boilers, corrosive condensates, and exhaust from black liquor evaporators. Weather Resistance. Indicators of the weather resistance of porcelain enamels are chemical durability, color stability, cleanability, and continuity of coating. Gloss and enamel texture do not necessarily affect weather resistance, but retention of gloss can be a criterion. Porcelain enamels have excellent resistance in atmospheric exposure, including corrosive industrial atmospheres, gases, smoke, salt spray, and seacoast exposures. Enamels formulated for acid resistance usually have better atmospheric-corrosion resistance than other types. These enamels have shown no appreciable change in appearance after 15 years of exposure. Waters. All porcelain enamels are completely resistant to water at room temperature. Resistance decreases at higher temperatures. Special porcelain enamels have been developed for hot- water storage tanks that can withstand continuous exposure to hot water for periods of 10 to 20 years. Natural waters, because of their varying compositions, have varying effects on porcelain enamels. Aerated water with a low dissolved solids content, for example, has been found to be more corrosive than hard water. Freezing and thawing cycles can cause some porcelain enamels to spall or disintegrate; however, properly formulated and applied enamels can withstand thousands of freezing and thawing cycles without failure. Porcelain enamels can withstand salt-spray tests (ASTM B 117) for days and even weeks without evidence of corrosion (Ref 1), and they can provide excellent service in intermittent or continuous exposure to seawater. Soils. Porcelain enamels formulated to withstand both acid and alkaline attack provide good service in soils. Acids. Resistance to acids varies widely, depending on composition and the application process used. The degree of attack by acids appears to depend less on the type of acid (with the exception of hydrofluoric) than on pH. Special formulations can provide excellent protection against aqueous solutions of most acids except hydrofluoric. The highest degree of acid resistance is obtained by sacrificing resistance to other media, such as alkalies. Figure 1 shows weight loss of a porcelain enamel in boiling mineral acids and in boiling water.

Fig. 1 Corrosion of a porcelain enamel in boiling water and boiling mineral acids. Source: Ref 2 Alkalies. Most porcelain enamels are unaffected by alkalies at room temperature. Special formulations provide alkali resistance to solutions with a maximum pH of 12 at temperatures to 100 °C (212 °F). Organics. Porcelain enamels are completely resistant to attack by common organic solvents, dyes, greases, and oils. Enamels are not dissolved by these materials and do not absorb them. Acid-resistant porcelain enamels are required for organic material that hydrolyzes upon contact with moisture to form acid solutions. High Temperatures. Porcelain enamels greatly reduce high-temperature oxidation of the base metal. This ability is largely due to the fact that the enamels themselves are fully oxidized and do not suffer further oxidation at elevated temperatures. They also form an effective barrier to the diffusion of oxygen into the metal. Protection depends on the temperature at which the enamel begins to soften and become more fluid. This temperature is normally about 220 °C (400 °F) below the firing temperature, but specially formulated enamels can provide oxidation protection to metals at temperatures to 1095 °C (2000 °F). Maximum service temperatures for porcelain enamels are shown in Table 6. Softening of the glassy matrix limits use to these exposure

temperatures. Softening releases gases remaining from reactions between the enamel and the metal substrate, producing random defects known as reboil. Table 6 Maximum service temperatures for porcelain enamels Service temperature °C °F 800 425 1000 540 1400 760 2000 1095

Limiting conditions

Usual limit for enamels maturing at about 815 °C (1500 °F) Maximum for enamels maturing at about 815 °C (1500 °F), without reboil Operating limit for special high-temperature enamels Refractory enamels useful for short periods for protection of stainless steels and special alloys

Source: Ref 3

References cited in this section 1. “Standard Method of Salt Spray (Fog) Testing,” B 117, Annual Book of ASTM Standards, American Society for Testing and Materials 2. C.L. Hackler and M. Dinulescu, Porcelain Enameled Flat Plate Heat Exchangers—Engineering and Application, Industrial Heat Exchangers, Proc. 1985 Symposium on Industrial Heat Exchanger Technology, A.J. Hayes et al., Ed., American Society for Metals, 1985 3. “Development of Porcelain Enamel Coatings,” Bulletin E-6, Porcelain Enamel Institute


Evaluation of Porcelain-Enameled Surfaces Specifications and quality control for porcelain-enamel coatings require the evaluation of a range of properties for the intended service of the porcelain-enameled product. Although material and process variables can be brought into approximate control using small test panels, process control is maintained by the evaluation of finished parts, even though some of the mechanical and chemical tests are destructive. Standard test procedures are available for many porcelain- enamel properties. Specific test methods for various properties are listed in Table 7. Some of these tests are discussed in the paragraphs that follow. Table 7 ASTM test methods for porcelain enamels Under jurisdiction of ASTM Porcelain Enamel Subcommittee B.08.12 Designation Title “Acid Resistance of Porcelain Enamels (Citric Acid Spot Test)” C 282 “Resistance of Porcelain Enameled Utensils to Boiling Acid” C 283 “Standard Test Methods for Sieve Analysis of Wet Milled and Dry Milled Porcelain Enamel” C 285 “Definitions of Terms Relating to Porcelain Enamel and Ceramic Metal Systems” C 286 “Standard Test Method for Adherence of Porcelain Enamel and Ceramic Coatings to Sheet C 313 Metal”

C 346(a) C 347

“Standard Test Method for 45° Specular Gloss of Ceramic Materials” “Standard Test Method for Reflectance, Reflectivity and Coefficient of Scatter of White Porcelain Enamels” “Standard Test Method for Fusion Flow of Porcelain Enamel Frits by the Flow Button Methods” C 374 “Classification of Water Used in Milling in Porcelain Enamel” C 375 “Standard Test Method for Thermal Shock Resistance of Porcelain-Enameled Utensils” C 385 (a) “Torsion Resistance of Laboratory Specimens of Porcelain Enameled Iron and Steel” C 409 “Standard Test Methods for Abrasion Resistance of Porcelain Enamels” C 448 “Standard Test Method for Continuity of Coatings in Glassed Steel Equipment by Electrical C 536 Testing” “Standard Test Method for Reliability of Glass Coatings on Glassed Steel Reaction Equipment C 537 by High Voltage” “Standard Test Method for Color Retention of Red, Orange and Yellow Porcelain Enamels” C 538 “Standard Test Method for Linear Thermal Expansion of Porcelain Enamel and Glaze Frits and C 539 Ceramic Whiteware Materials by the Interferometric Method” (a) “Standard Test Method for Image Gloss of Porcelain Enamel Surfaces” C 540 “Standard Test Method for Alkali Resistance of Porcelain Enamels” C 614 “Standard Test Method for Reboiling Tendency of Sheet Steel for Porcelain Enameling” C 632 “Standard Test Method for Adhesion or Cohesive Strength of Flame-Sprayed Coatings” C 633 “Practices for Production and Preparation of Gray Iron Castings for Porcelain Enameling” C 660 “Standard Test Method for Thickness of Diffusion Coating” C 664 “Standard Test Method for Weight Loss (Mass Loss) of Sheet Steel during Immersion in Sulfuric C 694 Acid Solution” “Standard Test Methods for Spalling Resistance of Porcelain Enameled Aluminum” C 703 “Standard Test Method for Nickel on Steel for Porcelain Enameling by Photometric Analysis” C 715 “Standard Test Method for Continuity of Porcelain Enamel Coatings” C 743 “Standard Test Method for Cleanability of Surface Finishes” C 756 “Standard Test Method for Yield Strength of Enameling Steels after Straining and Firing” C 774 “Standard Test Method for Nickel on Steel for Porcelain Enameling by X-Ray Emission C 810 Spectrometry” “Standard Test Method for Compressive Stress of Porcelain Enamels by Loaded-Beam Method” C 839 “Standard Test Method for Lead and Cadmium Release from Porcelain Enamel Surface” C 872 “Standard Test Method for Image Gloss of Porcelain Enamel Surfaces” C 988 (a) This test, while withdrawn from the latest ASTM list because of inability to obtain the required equipment, continues to be used by laboratories that possess the equipment. Adherence refers to the degree of attachment of enamel to the metal substrate. A number of tests regularly used in the industry provide adherence criteria, but none gives the force per unit area required to detach the enamel by tensile force normal to the interface. ASTM C 313 for porcelain enamel on steel deforms the metal and measures the area from which the porcelain enamel is removed. The indicator of adherence is the adherence index, which is the ratio of the porcelain enamel remaining in the deformed area to that in the same measured area prior to deformation. Enamels for cast iron pose a special problem because of the relatively greater thickness of the coating, the rigidity of the metal substrate, and the brittleness of the iron. Here, simple nonstandard impact tests are used. Resistance to spalling, a defect characterized by separation of porcelain enamel from the base metal without apparent external cause, is the indicator used to measure adherence of porcelain enamel on aluminum. Spalling can result from the use of improper alloys, improper enamel formulations, incorrect pretreatment of the base metal, faulty application, or unsatisfactory firing procedures. The most common test for spall resistance is ASTM C 703. Thickness. A number of specifications for products and applications require a specific thickness for porcelainenamel coating. The procedure used to measure the thickness depends on the type of base metal used. For porcelain-enamel products with a steel substrate, enamel thickness is measured according to ASTM D 1186. For porcelain-enamel products with an aluminum substrate, enamel thickness is measured according to ASTM E 376. (Neither of these tests is listed in Table 7 because they are not under the jurisdiction of ASTM Porcelain Enamel Subcommittee B.08.12.) In some cases, such as laboratory investigations, enamel thickness is measured according to ASTM C 664.

Resistance to Chipping. Relatively thick layers of porcelain enamel are fractured when subjected to severe bending or other substrate deformation. However, thin coatings of 125 μm (5 mils) or less that are well bonded to a relatively thin metal substrate (for example, 26-gage 0.4546 mm, or 0.0179 in., or less) can withstand the bending of the substrate to radii of curvature within its elastic limit and return to the original shape with little or no apparent damage. Chipping of typical porcelain enamel on sheet iron occurs at about the strain required for permanent deformation of the metal base. Abrasion Resistance. The resistance of porcelain enamel to various types of abrasion is measured by ASTM C 448, which consists of three parts. The first determines the resistance to surface abrasion of porcelain enamels for which the unabraded 45° specular gloss is more than 30 gloss units. The second determines the resistance to surface abrasion of porcelain enamels for which the unabraded 45° specular gloss is 30 gloss units or less. The third measures the resistance of porcelain enamels to subsurface abrasion. Acid Resistance. Porcelain enamels can be formulated to exhibit high resistance to all acids except hydrofluoric or concentrated phosphoric. This resistance to food acids and certain chemical cleaners is of particular importance to manufacturers of ranges and other kitchen and household appliances, as well as plumbing fixtures and various industrial and chemical processing products. The acid resistance of enamels under boiling conditions is important for certain uses, such as cookware. ASTM C 283 is used for evaluation. Similarly, acid resistance is a major consideration for predicting the weatherability of architectural porcelain-enamel components. Extensive testing under controlled-exposure conditions has shown a distinct correlation between acid resistance and weatherability. Here the test for determining acid resistance is ASTM C 282. Alkali Resistance. Home laundry equipment, dishwashers, and similar applications normally exposed to an alkaline environment at elevated temperatures require an alkali-resistant coating. The standard test for alkali resistance is ASTM C 614. Resistance to Organic Solvents. Porcelain enamels are inert to all common organic solvents; however, there is no standard procedure for determining this special performance characteristic. Resistance to Hot Water. Federal specification W-H-196J, “Heater Water, Electric and Gas Fired, Residential,” specifies a solubility test to determine the resistance of porcelain enamels to hot water. Weather Resistance. Actual weathering performance of porcelain-enamel panels has been documented in a series of on-site exposure tests carried out for up to 30 years by the National Bureau of Standards (now the National Institute of Standards and Technology) in cooperation with the Porcelain Enamel Institute. As mentioned previously, acid resistance as measured by ASTM C 282 is important. Weathering of porcelain enamel is evaluated in terms of the changes in gloss and color that occur during outside exposure. Multiyear weathering tests demonstrate that porcelain enamels have considerable inherent gloss and color stability. ASTM C 346 and ASTM D 2244 are used (the latter is not under B.08.12 jurisdiction). Spalling resistance of weathered porcelain enamel on aluminum is best ascertained by ASTM C 703. Torsion Resistance. In transit and during service, porcelain-enamel products may be subject to distortion through bending, twisting, or a combination of both. This can cause coating failures from fractures originating at the outer surface and normal to the tensional stress. ASTM C 409 evaluates the relative resistance to torsioninduced failure of laboratory test specimens for steel-substrate thicknesses of 24 to 12 gage. Thermal shock resistance of a porcelain-enamel surface varies inversely with the thickness of the enamel. It is also affected by the compressive stress present in the enamel at room temperature. Most porcelain enamels can be quenched in ice water from 205 °C (400 °F) without thermal shock failure. Some porcelain enamels specifically designed for resistance to thermal shock will not fail when quenched in ice water from 650 °C (1200 °F). ASTM C 385 is the standard for evaluating the durability of porcelain- enamel utensils when subjected to thermal shock. Continuity of Coatings. Ensuring continuity of coating after manufacture is important in porcelain-enamel (“glassed-steel”) applications where a prime purpose of the coating is to protect the substrate against corrosion. Test methods used to determine either discontinuity of coverage or potential discontinuity through too-thin coverage are ASTM C 536 and ASTM C 743. Both tests essentially involve the use of electrical probes of relatively high voltage to discern either discontinuities in the coating or insufficient coverage for coating integrity in service use.


Acknowledgment This article is adapted from “Porcelain Enamels,” Surface Engineering, Vol 5, ASM Handbook, ASM International, 1994, written by the Technical Publications Committee of the Porcelain Enamel Institute, Inc. who at that time included John C. Oliver, Consultant to the Porcelain Enamel Institute, Inc. (Project Chairman); Douglas D. Giese, G.E. Appliances (Co-Chairman); Jeffrey F. Wright, Ferro Corp. (Co-Chairman); Ronald L. Allen, Ronalco, Inc.; Terry Conrad, Amana/Raytheon Corp.; John K. Cook, Chi-Vit Corp.; Ted Furr, Bethlehem Steel Corp.; Andrew Gordon, U.S. Steel, USX Corp.; Bernard L. Hall, Jr., JennAir/Maytag Co.; Michael Horton, Industrial Heating and Finishing Co.; William Huyser, Mapes & Sprowl Steel, Ltd.; Robert J. Long, American Porcelain Enamel Co.; Anthony R. Mazzuca, Miles, Inc.; Herbert Oliveira, Nordson Corp.; Arnold Preban, Inland Steel Co.; Charles R. Rarey, LTV Steel Co.; Donald R. Sauder, Porcelain Enamel Institute, Inc.; Narayan M. Sedalia, Whirlpool Corp.; Jeffrey Sellins, Magic Chef/Maytag Co.; Lester N. Smith, Porcelain Consultants, Inc.; Larry L. Steele, Armco Steel Corp.; Ted J. Wolowicz, G.E. Appliances.


References 1. “Standard Method of Salt Spray (Fog) Testing,” B 117, Annual Book of ASTM Standards, American Society for Testing and Materials 2. C.L. Hackler and M. Dinulescu, Porcelain Enameled Flat Plate Heat Exchangers—Engineering and Application, Industrial Heat Exchangers, Proc. 1985 Symposium on Industrial Heat Exchanger Technology, A.J. Hayes et al., Ed., American Society for Metals, 1985 3. “Development of Porcelain Enamel Coatings,” Bulletin E-6, Porcelain Enamel Institute



“Alloy, Design and Fabrication Considerations for Porcelain Enameling Aluminum,” Bulletin PEI-801, Porcelain Enamel Institute, Norcross, GA “Ball Mill Wet Grinding of Slips for Porcelain Enameling,” Bulletin PEI-401, Porcelain Enamel Institute, Norcross, GA

• • • • • • • • • • • • • • • •

“Dipping and Flow Coating for Porcelain Enameling,” Bulletin PEI-502, Porcelain Enamel Institute, Norcross, GA “Electrostatic Porcelain Enamel Powder Application,” Bulletin PEI-501, Porcelain Enamel Institute, Norcross, GA “Enamel Preparation, Application and Firing for Porcelain Enameling Aluminum,” Bulletin PEI-803, Porcelain Enamel Institute, Norcross, GA Metallic and Inorganic Coatings, Vol 02.05, Annual Book of ASTM Standards, ASTM “1939 Exposure Test of Porcelain Enamels on Steel: 30-Year Inspection,” Building Science Series 38, National Bureau of Standards Porcelain Enamel Institute, Norcross, GA, www.porcelainenamel.com Porcelain Enameling, Materials, Finishing, and Coating, Vol 3, 4th ed., Tool and Manufacturing Engineers Handbook, Society of Manufacturing Engineers, 1985, p 23-1 to 23-10 Porcelain Enameling, Special Product Supplement, Appliance Mag., March 1988 and April 1990 “Preparation of Sheet Steel for Porcelain Enameling,” Bulletin PEI-301, Porcelain Enamel Institute, Norcross, GA “Pretreatment of Alloys for Porcelain Enameling Aluminum,” Bulletin PEI-802, Porcelain Enamel Institute, Norcross, GA “Properties of Porcelain Enamel,” Data Bulletins PEI 501, 502, 503, 504, and 505, Porcelain Enamel Institute, Norcross, GA “Quality Control Procedures for Porcelain Enameling Aluminum,” Bulletin PEI-804, Porcelain Enamel Institute, Norcross, GA “Weather Resistance of Porcelain Enamels: 15-Year Inspection of the 1956 Exposure Test,” Building Science Series 50, National Bureau of Standards “Weathering of Porcelain Enamels on Aluminum: 12-Year Inspection, Exposure period 1964–1976,” Department of Ceramic Engineering, University of Illinois at Urbana- Champaign, 1977 “Wet Spraying for Porcelain Enameling,” Bulletin PEI-503, Porcelain Enamel Institute, Norcross, GA J.F. Wright, C.G. Bergeron, and J.C. Oliver, Porcelain Enamel, Ceramics and Glasses, Vol 4, Engineered Materials Handbook, ASM International, 1991, p 937–942

Chemical-Setting Ceramic Linings, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 755–758

Chemical-Setting Ceramic Linings Introduction INORGANIC CHEMICAL-SETTING CERAMIC LININGS have become one of the most widely used construction materials in designing protective linings for industrial installations in which high temperatures, aggressive corrosive media, and complicated substrate geometry exist, such as floors, trenches, sumps, reaction vessels, tanks, scrubbers, ducts, chimneys, and other air pollution control equipment. They are used in various industries, including power, steel- and metalworking, chemical, pulp and paper, refinery, waste treatment, and mining. Inorganic monolithic linings have proved themselves in these industries because of their chemical resistance to both high and low concentrations of strong acids and solvents, thermal insulation that protects the substrates from extremely high temperatures, temperature resistance to 870 °C (1600 °F), good compressive and flexural strength for environments in which stress and strain are factors, and abrasion resistance. Monolithic linings can be applied by cast or gunite (shotcreting) methods over old and new steel or concrete as well as brick and mortar masonry. This article discusses the function of monolithic linings, the use of these materials, the types of applications in which these materials can be successfully used, and the limitations of these linings.

Chemical-Setting Ceramic Linings, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 755–758 Chemical-Setting Ceramic Linings

History of Chemical-Setting Silicates The progress of silicate cement development has changed over the years with growing technology to meet industry needs. The first silicate cements were composed of a sodium silicate (Na2SiO3) liquid and a combination of fillers, such as silica flour, clay, silica aggregates, and barytes, and were formulated as chemical-resistant mortars for use in ambient or high-temperature acid lining construction. This type of silicate cement was very slow in setting and had to be exposed to the open air or heat cured, which created construction delays. Another problem was that these cements were not water resistant. These problems were resolved by the use of acid washing, which helped set the cements and make them water resistant. In the early 1920s, Na2SiO3 mortars were introduced that used an acid catalyst to insolubilize the silica gel; this produced a mortar that cured faster and was water resistant. The physical properties of the silicate mortar were unchanged; that is, the acid, temperature, and solvent resistance were maintained. Because the acid-catalyzed mortar was chemically activated on mixing, it could set within 24 to 48 h. This was a distinct advantage for the construction industry, because it allowed brick to be laid continuously without concern over the mortar being squeezed out of the joints or the brick sliding out of line. Typical setting agents used are ethyl acetate (C 4H8O2), zinc oxide (ZnO), sodium fluorosilicate (Na2SiF6), glyceryl diacetate, formamide (CH3ON), metallic polyphosphates, and other amides or amines. It was later determined that when Na2SiO3 mortars were exposed to sulfuric acid (H2SO4), the reaction product was sodium sulfate (Na2SO4), which is a salt that expands and grows through hydration. This sodium salt can pick up as much as 10 mol of water of crystallization with a resultant size increase of 150%, thus creating internal stresses in the structure in which it was used. As a result, monolithic linings and brick surfaces would sometimes crack or spall. As time passed, industry required silicate technologies that would meet changing applications and structural designs. By the 1950s, potassium silicate (K2SiO3) cements were introduced using methods of insolubilizing the silicate. The K2SiO3 cements possess enhanced corrosion resistance, particularly in H2SO4 environments. Potassium silicate cements react with H2SO4 to form potassium sulfate (K2SO4), which is not a growth salt. This eliminates the problem of internal stresses when the mortar is exposed to H2SO4 solutions. The introduction of K2SiO3 cements was a positive step in producing a material that was more suited for application technology and corrosive environments. The K2SiO3 cements offer improved physical properties, which are particularly beneficial in monolithic applications. Furthermore, K2SiO3 cements provide better workability with less tackiness and longer working times than Na2SiO3 cements. Additional advantages include greater resistance to strong acid solutions and sulfation, more refractoriness, and no efflorescence as with Na2SiO3 cements. By this time, Na2SiO3 and K2SiO3 materials were not only being used as mortars but were also being introduced in new areas of monolithic application as both castable and gunite grades. Modified silicates were developed in the 1980s. Modified silicates are manufactured by the addition of a powder form of Na2SiO3 in conjunction with a proprietary ingredient added directly to the powder fillers. This process eliminates the need to have both powder and liquid on the job site; water is the only required addition. Although these products have been commercially available, high cost and potential for sulfation have limited their use. The modified silicates do not have the chemical properties of the other silicates. The only advantage appears to be their simplicity of mixing.


Advantages and Disadvantages The Na2SiO3 and K2SiO3 cements were developed because a material was needed that was acid resistant in dilute-to-strong concentrations and that offered higher temperature resistance than chemical-resistant organic materials. The development of these silicate-base acidproof cements resulted in many advantageous characteristics, such as a 100% K2SiO3-bonded system that had resistance to most solvents and acids over a pH range of 0 to 7; that was water and vapor resistant without special treatment; and that could withstand all concentrations of H2SO4, nitric acid (HNO3), hydrochloric acid (HCl), and phosphoric acid (H3PO4). In addition, the Na2SiO3- and K2SiO3-bonded cements can be applied over damp acid-attacked concrete or brick surfaces as long as these substrates have acid pH surfaces, and they cure chemically within 36 h, thus decreasing construction delays. These cements can be applied by gunite or cast methods. Detailed information on guniting (shotcreting) can be found in American Concrete Institute publication ACI 547R on refractory concrete. Certain disadvantages were encountered during the development of the acid-resistant silicate cements. It was common knowledge that the silicate cements were not resistant to alkalis, hydrofluoric acid (HF), and fluoride salts. As previously mentioned, the Na2SiO3 cements formed a growth salt when exposed to H2SO4 that put undue internal stresses on the structure of the material. Inorganic monolithic linings also have a certain amount of permeability compared to organic surfacing materials. Over time, acid can penetrate the lining and eventually reach the surface of the substrate. This problem is now being combated by using a dual-lining system, which includes a chemically resistant elastomeric membrane applied to the surface of the substrate.


Current Technologies Monolithic and Membrane Dual Linings. Organic linings or coatings have been used in stacks and ducts, but they will usually deteriorate if temperatures exceed 260 °C (500 °F) in high- sulfur gases. There are many types of lining materials for stacks and ducts; however, because of unpredicted environmental conditions within a stack or duct, it may be difficult to select the best candidate material. Stacks that are designed with heat recovery units and/or scrubbers, combined with one or more flues inside the shell that can control the temperature, should be considered for dual-lining applications. In recent years, the trend toward using the superior technology of dual linings has emerged and is being recommended where corrosion problems occur throughout industry. Figure 1 illustrates the design of a typical membrane/ monolithic system in the chemical industry.

Fig. 1 Schematic of a chemical-resistant dual-lining system that provides double protection to the substrate in the form of a flexible membrane and a rigid surface layer. The flexible corrosion-resistant membrane is applied in direct contact with steel or concrete substrates. It is then covered by the monolithic cement lining, which provides protection over a broad pH range as well as against high temperatures. There are numerous reasons for this new construction process. Condensation occurs not only on the face of the acid-resistant lining but can also penetrate and condense on the substrate to be protected. Although acidproof monolithic linings offer the proper chemical resistance, they are inherently inelastic, or brittle. In time, monolithic linings may tend to crack and absorb acids or acid gas condensate; therefore, it is advantageous to have a backup membrane. In many applications, the coefficient of thermal expansion of the monolithic lining may not match that of the substrate. Therefore, a flexible membrane helps accommodate stresses resulting from these differences in thermal expansion as well as other mechanically induced stresses. For many applications, the monolithic lining should not be bonded to the substrate, and the membrane acts as a bond barrier. The concept of placing an impervious membrane between the substrate and inorganic linings has become a recommended practice. Some corrosion-resistant gunite manufacturers have installation specifications for use with membranes. Membrane choices include asphaltics, resins, and synthetic elastomers. The membrane selected should resist the maximum acid concentrations and temperature expected. Organic linings can fail by disbonding, swelling, abrasion, and blistering from high temperature. Protection of an organic membrane with an inorganic barrier will minimize thermal exposure and mechanical abrasion to the membrane as well as the corrosive media that reach the membrane. An organic membrane compensates for some of the shortcomings of inorganic linings, such as cracking or spalling due to mechanical stresses from shrinkage, induced stresses, insufficient thermal expansion allowances, the anchoring system, vibration, and thermal or other stresses. To provide an effective barrier, inorganic linings are applied much thicker than typical organic coatings. They often contain fillers to act as reinforcement and to decrease shrinkage. Fillers are also incorporated to provide wear resistance for abrasive conditions. The installation procedure for a dual-lining system is basically the same as if each component were being applied separately. These procedures are established for both organic coatings and inorganic monolithic linings. The condition of the working area and preparation of the surfaces follow standard practices used throughout the industry. Anchors are usually installed on the steel or concrete substrates before the organic membrane is applied. Lightweight Insulating Materials. Industry has recently seen the development and application of improved lower-density, lightweight, insulating corrosion-resistant lining materials. The evolution of these products has followed the need to accommodate load limitations of structural supports and improved thermal protection of membrane coatings where substrate temperatures can range from 40 to 150 °C (100 to 300 °F) and even higher if process limits or gas cleanup systems experience frequent excursion or bypass conditions. Membranes

require adequate thermal protection to retain physical and chemical properties throughout their service life. Because the membrane will also coat the studs, the inorganic coating must be thick enough to protect the membrane on the highest point of the studs. Therefore, the inorganic lining material must not only protect the membrane from physical abuse but must also protect it from thermal deterioration. When membranes are not used, condensation of chemicals from hot process streams should be designed to occur within a lining and not at the substrate. However, when membranes are used, the condensation can occur at the membrane, thus possibly reducing the required lining thickness. Therefore, inorganic silicate lining materials with lower thermal conductivity and density have been developed to meet these needs. Application Procedures. Through the use of current technology, such as the dual-lining system and the availability of improved application techniques, corrosion protection practices have been greatly enhanced. Observations of successes and failures in various applications have led to the development of the following parameters, which must be addressed to specify the proper material. In any application, the primary consideration is the exposure environment. This includes the types and concentrations of chemicals, physical abuse, and temperatures. In this evaluation, it is important to remember that both the silicate monolithic lining and the membrane must resist the corrosive environment. The next consideration is the thickness of the monolithic lining. In high-temperature environments, the monolithic lining must be able to decrease the temperature through its thickness to a level that does not burn out the membrane applied to the substrate or the anchoring system. Once the materials that can resist the corrosive environment have been selected, the physical structure must be considered. Steel. When a steel substrate is being used, sources of variation in electrical potential on the metal surfaces must be considered. These include mill scale, metal impurities, localized strains on parts of the metal, junctures of dissimilar metals, movement of electrolytes, weld joints, concentration cells, and externally imposed currents. Some or all of these differences in potential can occur on the surface and must be considered in the recommendation of a protective coating. The protective coating must provide the following properties to protect the steel: • • • • • • •

Physical protection to surface Prevention of concentration cells Dielectric properties Lower surface temperatures Prevention of electrolyte flow Deprivation of free oxygen from the surface Stress relief

Stress relief is an important factor in the design of a system. If a system does not compensate for relief in areas of high stresses, cracking will occur that can severely damage the entire system. Stresses on the system of major concern could result from vibration, unsupported surfaces, changes in planes, and welds. The following five steps are application procedures that should be followed to ensure proper installation of a dual-lining system. First, all areas of high stress that result in movement of the steel structure, such as oil canning or vibration, must be externally supported. Second, the anchoring system should then be applied to the substrate. Anchors, such as V-type or longhorn studs, should be installed on the proper centerlines in a diamond or staggered pattern with the tines randomly oriented. Third, the steel surface must be stripped of all oil and grease by chemical cleaning. The metal should then be sandblasted to an SP 10 Near-White Blast (SSPC: The Society for Protective Coatings designation) with a nominal 64 μm (2.5 mils) profile. Welds should be ground to a smooth, rounded radius with no sharp edges before blasting. Fourth, the appropriate membrane should be applied, ensuring that all of the studs are completely coated and that the system is free of pin holes. The final step is to apply the monolithic lining. The two available methods of application should be considered. A castable material will shrink more than the gunitable material; therefore, expansion joints must be planned into the system in areas exceeding 6 by 6 m (20 by 20 ft) or where changes in planes will cause stresses to occur. A gunitable material will shrink less and therefore does not require expansion joints. Figure 2 illustrates the gunite application of K2SiO3 cement over steel ductwork in a fossil fuel power plant.

Fig. 2 Gunite application of 50 mm (2 in.) of K2SiO3 cement over steel ductwork in a coal-fired power plant Concrete. In applications over concrete, as with steel, the first considerations are the corrosive environment and the thickness of the material. Once the materials have been selected, the physical structure must be considered. When the substrate is concrete, factors different from those associated with steel must be considered. Inorganic monolithics can be applied over old or new concrete surfaces. When concrete substrates are being used, the following questions should be answered: • • • • • •

Is the concrete old acid-attacked concrete or new concrete? Is the concrete capable of supporting loads? Is the structure above grade or below grade? Does the concrete need to be waterproofed from the outside? Does the concrete structure have existing cracks? Is the concrete surface a high or low pH?

A combination of many of these items may apply and should be considered in the recommendation of the proper system. The following are the criteria for a concrete surface before a system may be applied. When working with old acid-attacked concrete, all loose and deteriorated concrete should be removed. The surface should be firm, hard, and at a minimum pH of 5, and all surfaces should be brought back to grade and the slopes reestablished. When working with new concrete, the concrete should be firm and sound. All structural cracks should be repaired, and it should be cured for a minimum of 7 days. Slopes should be a minimum 3.2 mm (0.125 in.) to a maximum 6.4 mm (0.25 in.) per linear foot and should have attained a minimum compressive strength of 21 MPa (3000 psi). The following five steps are application procedures that should be followed to ensure proper installation of a dual-lining system. First, the concrete should meet the standards outlined previously. Second, the concrete should be cleaned of all oil, grease, and form release compounds by chemical cleaning, waterblasting, or scarifying. Third, the anchoring system should be placed at the specified centerline distance in a diamondshaped pattern with a random orientation of anchor tines. The anchoring system should consist of V-type or longhorn studs. Fourth, the membrane should be applied to the recommended thickness, ensuring that all studs are completely coated and that the coating is free of pinholes. The final step is to apply the monolithic. The castable and the gunitable methods should both be considered. As stated previously, when applying by the gunite method there is less shrinkage; therefore, there is no need for expansion joints in the monolithic. However, the lining will require control joints in the normal manner. Expansion joints are recommended with cast applications and should be installed at the perimeter of floor areas,

equipment pads, and floor trenches; around floor drains, column bases, and protrusions; at changes in planes; and over expansion joints in concrete slabs not to exceed 6 m (20 ft) on centerlines. Brick. A monolithic lining can be applied over acid-attacked masonry construction. The first step is the surface preparation of the brick. Fly ash and other contaminants should be removed by sandblasting or waterblasting. All attacked or unsound mortar should be removed from the joints. Mortar joints should be cleaned to a depth of at least 13 mm (0.5 in.) to provide support for the monolithic lining. If the joints cannot be cleaned to this depth, it is necessary to set longhorn or V-type anchors in the joints. The preferred method of application in this case is to gunite the material. Guniting allows the material to be applied under pressure, thus enabling the material to be packed into the 13 mm (0.5 in.) open joints and allowing the system to be supported by the studs or keying into the joints.


Corrosion Resistance Monolithic linings are among the fastest growing products in the corrosion protection field because of their resistance to a wider range of corrosive media than most other construction materials. The following industrial applications illustrate the corrosion resistance and some uses of monolithic linings. Wastewater treatment systems in the southern United States have experienced corrosion problems in areas of high temperatures, high humidity, and slow drainage. Consequently, millions of dollars have been spent in replacing the corroded concrete infrastructures in these wastewater treatment systems in wet wells, grit chambers, sewer lines, and aeration basins. This corrosion went unnoticed throughout the years because the corrosion mechanism was not understood. Most of the corrosion was not caused by discharged chemicals but was microbiologically induced. The concrete infrastructures are seldom completely filled, and the combination of high temperatures and high humidity creates a perfect breeding ground for bacteria, fungi, algae, and gaseous products from decomposing sewage. Carbon dioxide (CO2), combined with hydrogen sulfide (H2S) from the slightly warm and fermenting sewage, reacts with the damp concrete surfaces to form carbonates and calcium sulfate (CaSO4). As this process proceeds, the pH becomes more acidic. Bacterial action results in the production of elemental sulfur from various sulfate reactions, with subsequent oxidation of elemental sulfur to H2SO4, which the bacteria secrete. The H2SO4 then directly attacks the underlying portland cement concrete substrate and causes destruction of the infrastructure. There are documented cases of microbiologically induced corrosion having proceeded to the point at which structures collapsed in as few as 6 years or required major rehabilitation in as few as 4 years. A proven method of rehabilitating deteriorated concrete wastewater structures involves applying a urethanebased membrane to the properly prepared substrate, followed by the pneumatic application of a 100% K2SiO3bonded acid-resistant concrete. The urethanebased membrane is a two-component material with a high solids content and very low permeability. To ensure pinhole-free coatings, the membrane is applied in two or more coats to a wet film thickness ranging from 1.6 to 3.3 mm (63 to 130 mils). These membranes are applied by airless spray with a spray tip pressure of approximately 28 MPa (4000 psi). The high pressure further helps to ensure pinhole-free application of the membrane. The substrate must be dry enough to accommodate the urethane membrane without reacting with the polyisocyanate catalysts usually used. Therefore, it is sometimes desirable to apply a moisture-tolerant primer, such as a suitable formulated epoxy, to the damp concrete or to apply a fresh layer of concrete first. Following the proper application and cure of the urethane membrane, a veneer of a 100% K2SiO3 acidproof cement is applied over the membrane and anchors. The chemical industry can be the most difficult of all industries in which to design a protective lining. The conditions within one plant can be staggering; there may be acids, alkalies, and solvents, along with extremely high temperatures and physical abuse. Potassium-silicate- base cements satisfy most of these conditions while maintaining economy. The combination of a wide range of chemical resistance, high physical strength, and refractoriness makes these cements leading candidates for many chemical plant applications.

For example, at a major Midwest chemical plant, the action of HCl by-products and ambient moisture on floors, digester tank supports, trenches, and sumps in NiCl2- and NiSO4-processing units produced very corrosive conditions. Restoration of the existing structures was accomplished with a K2SiO3 cement and an appropriate membrane. In the NiCl2 unit, the existing concrete floor was chipped out, and new concrete was poured; the steel beam supports for the digester tanks were also encapsulated with regular concrete. Both floor and digester foundations then received a trowel-applied asphaltic membrane. The membrane provided a flexible acid-, water-, and alkali-resistant barrier between the concrete substrate and the monolithic surfacing. The trowelapplied membrane system is applied at room temperature and offers the chemical resistance of hot-applied membranes for those areas where a hot membrane may not be practical or advisable. The next step was to apply a K2SiO3 monolithic lining by casting into forms. The lining was 38 mm (1. 5 in.) thick over foot traffic areas near the digesters; 38 to 50 mm (1.5 to 2 in.) thick around the encapsulated digester foundation, creating a collar; and 50 mm (2 in.) thick in forklift truck aisles. Finally, at the interface of the digester tank wall and the collar, an elastomeric urethane-based compound was used as a permanently flexible expansion joint. In the NiSO4-processing unit, the problem areas included trenches running among process equipment in two different buildings. In addition, the various collection sumps into which these trenches drained required restoration. As was done in the NiCl2 area, the deteriorated concrete was reconstructed with new concrete. The asphaltic membrane was trowel applied. A K2SiO4 monolithic lining was cast over the membrane to 25 mm (1 in.) in the trenches, 25 to 38 mm (1 to 1.5 in.) in the sumps, and 38 mm (1.5 in.) on the floors, and the elastomeric urethane expansion joint compound was applied. The service conditions for the two processing areas included a pH of 1 to 2 due to concentrations of HCl up to 32%, water, temperatures to 95 °C (200 °F), and vibrations from imposed loads as great as 3630 kg (8000 lbs) in the truck aisles. These applications were done by the casting method, but they could have easily been done by the gunite method.


Acknowledgment This article has been adapted from G.D. Maloney, Chemical-Setting Ceramic Linings, in Corrosion, Volume 13, ASM Handbook (formerly Metals Handbook, 9th ed.), ASM International, 1987, p 453–455.



W.W. Carpenter, Ceramic Coatings and Linings, Surface Engineering, Vol 5, ASM Handbook, ASM International, 1994, p 469–481 J.R. Davis, Ed., Surface Engineering for Corrosion and Wear Resistance, ASM International, 2001 R.A. Eppler, Ceramic Coatings, Ceramics and Glasses, Vol 4, Engineered Materials Handbook, ASM International, 1991, p 953–958

• • • • •

A.O. Kaeding, “Corrosion Resistant Pipe Liners—Polymer Concrete”, SPB9-02, American Concrete Institute, 1985 J.S. Reed, Introduction to the Principles of Ceramic Processing, John Wiley & Sons, 1988 “Refractory Concrete: State-of-the-Art Report,” ACI 547R-97, American Concrete Institute, 1997 P.A. Schweitzer, Corrosion-Resistant Linings and Coatings, Marcel Dekker, 2001 K.H. Stern, Ed., Metallurgical and Ceramic Protective Coatings, Chapman & Hall, 1996

CVD and PVD Coatings, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 759–762

CVD and PVD Coatings Introduction VAPOR-DEPOSITED COATINGS modify the surface properties of materials. They are widely used and have been particularly successful in improving such mechanical properties as wear, friction, and hardness in cutting applications. Their use as corrosion-resistant coatings is also becoming widespread.

CVD and PVD Coatings, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 759–762 CVD and PVD Coatings

Deposition Processes The vapor-deposition processes fall into two major categories—physical vapor deposition (PVD) and chemical vapor deposition (CVD)— which in turn comprise various subcategories. Although these divisions often appear confusing and arbitrary, each has its own well-defined technical and economic advantages and disadvantages. This is also true of the materials used in vapor deposition. Therefore, the need for a thorough understanding of both processes and materials is needed for the selection of the optimal combination. The major deposition processes that are reviewed in this article are sputtering, evaporation, ion plating, and CVD. These processes all have one characteristic in common: the deposition species are transferred and deposited onto the substrate in the form of individual atoms or molecules. In this respect, they are fundamentally different from particulate- or liquid-deposition processes, such as thermal spraying or electroplating, and ion-beam processes where ions are implanted beneath the surface of the material (see the articles “Thermal Spray Coatings,” “Electroplated Coatings,” and “Surface Modification Using Energy Beams” in this Volume). A primary benefit of vapor deposition is that it is essentially nondamaging to the environment, because there are no solvents or chemicals to dispose of. With the current awareness of environmental, safety, and health considerations, this crucial factor favors vapor deposition over other processes that may be more harmful to the environment and health. Sputtering is the principal PVD process. It involves the transport of a material from a source (target) to a substrate by bombardment of the target by gas ions that have been accelerated by a high voltage. Atoms from the target are ejected by momentum transfer between the incident ions and the target. These ejected particles move across the vacuum chamber to be deposited on the substrate. Figure 1 is a schematic of the sputtering process. In its simplest form, the process occurs in an inert (noble) gas at low pressure (0.13 to 13 Pa, or 1 × 10-3 to 0.1 torr); in most cases, the gas is argon. Argon has higher mass than other noble gases, such as neon or helium, and is easier to ionize. Higher mass gives a higher sputtering

yield, especially if the mass of the bombarding particle is the same order of magnitude or greater than the mass of the target atom. Other gases, such as oxygen or nitrogen, can also be used, but they may react chemically with the target. The sputtering process begins when an electric discharge is produced and the argon becomes ionized. The low-pressure electric discharge is known as glow discharge, and the ionized gas is termed plasma.

Fig. 1 Basic sputtering process The argon ions hit the solid target, which is the source of the coating material (not to be confused with the substrate, which is the item to be coated). The target is negatively biased and therefore attracts the positively charged argon ions, which are accelerated in the glow discharge. This attraction of the ions to the target (also known as bombardment) causes the target to sputter, which means that material is dislodged from the target surface because of momentum energy exchange. The higher the energy of the bombarding ions, the higher the rate of material dislodgement. Sputtering developed very rapidly in the 1970s in the semiconductor industry, in which the technique is essential for mass production. Sputtering is used in corrosion-abatement applications such as high-chromium and nickel-chromium alloy coatings, MCrAlY (where M stands for nickel, cobalt, or iron or a combination of them), and some polymers. Titanium nitride, other nitrides, tungsten, molybdenum, carbides, borides, and diamondlike carbon are sputtered for wear resistance (Ref 1). For the deposition of thin metal films (in the 0.1 μm, or 4 μin., range), sputtering is the best technique. Deposition is feasible in a controllable manner for both compound and elemental targets (the source of the coating material). Adhesion is good and can be further improved by sputter cleaning the substrate or by bias sputtering. Large equipment is readily available, and sputtering has become a highly automated process. The quality, structure, and uniformity of the deposit are excellent. The disadvantages of sputtering include its thickness limitation, the need for a vacuum, and high cost. It is also a line-of- sight process that requires special fixtures and shaped targets to coat uniformly; furthermore, the coating of compound curves, recesses, and holes is difficult. The substrate being processed must be capable of withstanding processing temperatures of 260 to 540 °C (500 to 1000 °F). Evaporation was the first PVD process used on an industrial basis for the aluminum metallization of plastics and glass for decorative purposes. Originally the most widely used process, evaporation has now been overtaken by sputtering. The metallization process is relatively simple and well adapted to mass production. The basic evaporation process involves the transfer of material to form a coating by physical means alone, essentially vaporization. Practically all metals can be evaporated, making evaporation a universal process. The source material may be heated by bombardment with an electron beam or heated by thermal conduction. Like sputtering, it is a line-of-sight process; therefore, to achieve efficiency of coverage and uniformity it is necessary to use multiple evaporation sources and to rotate or move the substrate uniformly to expose all areas.

Unlike other vapor-deposition processes, evaporation is a low-energy process, with particle energy averaging 0.2 to 0.3 eV as compared to 10 to 40 eV for sputtering, 200 eV or more for ion plating, and 100 keV for ion implantation. The term “low energy” means that adhesion of the particle to the substrate may be marginal. In addition, the temperature of the substrate is not increased by the deposition, which means there is little or no diffusion. Because no strong physical or chemical action takes place at the interface, particular care must be taken to remove all impurities in the chamber and to have a very clean substrate. The basic process (with an electron-beam source) is shown in Fig. 2.

Fig. 2 Basic evaporation process Very large evaporation equipment, which was developed for the optical industry, is commercially available. Coatings of several microns in thickness can be rapidly deposited. The disadvantages of this method are nonuniform coverage (limited to line of sight) and relatively low adhesion unless glow-discharge cleaning is used (glow-discharge cleaning removes surface atoms from the substrate). Evaporation is very successful in applications in which adhesion and good structure are not critical, such as decorative and optical uses. Aluminum is the principal evaporation material; chromium and stainless steel are also widely used. Evaporation is not generally used for critical corrosion-control applications. Ion plating is a generic term for the hybrid process based on the evaporation (or sputtering) mechanism coupled with a flux of high-energy particles (Fig. 3) imparting a large increase in the energy of the deposition species, which is atomistic.

Fig. 3 Basic ion plating process Ion-beam-assisted deposition (IBAD) and ion- beam-enhanced deposition (IBED) refer to the process wherein evaporated atoms produced by physical vapor deposition (PVD) are simultaneously struck by an independently generated flux of ions (Fig. 4) (Ref 2). The extra energy imparted to the deposited atoms causes atomic displacements at the surface and in the bulk, as well as enhanced migration of atoms along the surface. These resulting atomic motions are responsible for improved film properties—including better adhesion and cohesion of the film, modified residual stress, and higher density— when compared with similar films prepared by PVD without ion bombardment. When the ion beam or the evaporant is a reactive species, compounds such as refractory silicon nitride (Si3N4) can be synthesized at very low temperatures. Furthermore, adjustment of the ratio of reactive ions to atoms arriving at the substrate surface allows adjustment of the stoichiometry of solid solutions.

Fig. 4 Ion-beam-assisted deposition (IBAD) It is generally accepted that plasma plays the same part as high temperature in improving the adhesion and increasing the reactivity of the coating; an ion passing through a potential of 100 eV acquires an equivalent temperature of 106 K. While the PVD process uses temperatures of approximately 538 °C (1000 °F), the IBED process deposits hardcoats below 93 °C (200 °F) (Ref 3).

The simplest forms of ion plating are ion nitriding and ion carburizing, in which nitrogen or carbon ions react with a steel substrate to form nitrides or carbides. If the deposition species is ionized in a reactive gas, a compound is formed with some of the gas atoms, and deposition of the compound occurs. Ion plating is widely used and is gradually replacing pack cementation or diffusion. With ion implantation, the objective is to modify the properties of the surface, such as corrosion behavior, by embedding atoms from a beam of ionized particles into the near surface of the target. Because ion plating is a glow-discharge process in which a plasma is formed, it is necessary to have at least a threshold gas pressure to maintain the plasma. This is different from standard nonionized evaporation, in which the minimum pressure obtainable is preferred. The net effect of this higher gas pressure is to increase the number of collisions with gas molecules and therefore produce more scattering of the deposition species, which in turn diminishes the importance of the line-of-sight principle and improves coverage. The deposited species in ion plating have higher energy than those of evaporation or sputtering processes; this results in improved adhesion as well as a deposit with improved structure and fewer imperfections. In addition, if a sputtering system is used, it is easy to sputter the surface of the substrate to clean it before deposition. However, the ion equipment is more complicated and therefore more expensive than either evaporation or sputtering equipment, and the process is more exacting. Chemical vapor deposition is a well-established deposition process that is used extensively in the semiconductor and cutting-tool industries. It has the unique ability to deposit thick, dense, high-quality films (up to 6.4 mm, or 0.25 in., thick or more) at a cost that is generally lower than that of PVD processes. The high temperature, about 1000 °C (1830 °F), necessary for CVD promotes good adhesion, but also considerably restricts the type of substrate that can be coated to ceramics, refractory metals, and special alloys. In most cases, steel will require heat treatment in a vacuum or protected atmosphere after CVD coating and will often change dimensions sufficiently to require postdeposition machining if close tolerances are required. Another constraint resulting from the high temperature of deposition is the stress in the deposit due to the difference in thermal expansion between substrate and deposit. These stresses may be sufficient to cause cracking, spalling, and loss of adhesion. The technique of plasma CVD can be used at much lower temperatures in, for example, semiconductor applications, but the resulting films tend to incorporate impurities. The materials that can be easily deposited by CVD are more limited than with PVD. Chemical vapor deposition is particularly useful for depositing compounds and refractory metals and lends itself very well to corrosion applications. Figure 5 is a schematic of thermal CVD equipment. The reactive gases are metered in the mixing chamber. They then pass into the deposition chamber and contact the heated part to be coated (sample); at this point, they react and deposition occurs. By-products are removed through the vacuum system.

Fig. 5 Thermal CVD reactor

Chemical vapor deposition reactions can occur over a range of pressures. Low pressure increases boundarylayer diffusion and uniformity, often at the expense of deposition efficiency. Each application and each reaction must be analyzed to determine the optimal conditions for deposition.

References cited in this section 1. J.R. Davis, Ed., Surface Engineering for Corrosion and Wear Resistance, ASM International, 2001, p 175 2. G.K. Hubler and J.K. Hirvonen, Ion-Beam- Assisted Deposition, Surface Engineering, Vol 5, ASM Handbook, ASM International, 1994, p 593 3. A.H. Deutchman and R.J. Partyka, Comparison of the Properties of PVD and IBED Hardcoats (TiN and Cr2N), First ASM International Surface Engineering Congress and 13th IFHTSE Congress, Oct 2002 (Columbus, OH), ASM International 2003, p 138


Coating Materials and Applications The list of materials that can be vapor deposited is extensive and covers almost any coating requirement. The most important materials are given in Table 1. Aluminum and aluminum alloys are among the most widely used deposition materials and are gradually replacing cadmium in many corrosion applications. Sputtered chromium and stainless steel are also making great inroads in corrosion applications. A promising material is titanium nitride. It is hard, very stable, and refractory, and it has good lubricating characteristics and a pleasing gold appearance. It is widely used in wear applications, such as cutting tools, and in decorative applications. It is usually applied by sputtering or ion plating, and it holds potential for corrosion protection. Table 1 Some corrosion-resistant vapordeposited materials Material Aluminum Chromium Nickel Rhenium Tantalum Titanium Tungsten Boron Carbon Titanium diboride Silicon carbide Titanium carbide Tungsten carbide Boron nitride Silicon nitride Titanium nitride

Deposition process Evaporation Sputtering CVD, sputtering CVD Sputtering, CVD Evaporation, sputtering CVD, sputtering CVD CVD, ion plating Evaporation, CVD Sputtering, CVD Sputtering, evaporation, CVD Sputtering, CVD CVD Sputtering, ion plating CVD, sputtering, ion plating

Sputtering, ion plating Aluminum oxide Sputtering, ion plating Molybdenum silicide Nickel-chromium alloys Sputtering Evaporation Gold alloys CVD Borazon Evaporation, sputtering MCrAlY Evaporation, sputtering Bronze The properties of these materials vary widely and reflect the complexity of selecting a corrosion-resistant coating for a given substrate or a given environment and the need to conduct a thorough analysis of the problem. This is especially true when cost is the primary factor. For example, gold is an ideal material for many corrosion applications, but would of course be cost prohibitive in most cases. Before selecting a material and process for a corrosion-resistance application, the following considerations should be reviewed: • • • • •

Availability and cost of the coating material Cost, size, complexity, and availability of the deposition equipment Resistance to the chemical environment of the coating material Compatibility with the substrate, especially thermal expansion mismatch In special applications, such as aerospace, high-temperature resistance, radiation resistance, and highvacuum stability

Chemical vapor deposited and PVD coatings have the advantage that the film (if thick enough) is essentially pore free and fully dense. Penetration by moisture and gases to the substrate is greatly reduced, if not eliminated. Despite this, the role played by CVD and PVD coatings in corrosion protection remains small. Some current applications are discussed in the paragraphs that follow. Coating for Graphite. Graphite is perhaps the best material for applications requiring high strength and low density at very high temperatures. It is used as a hot-pressed material or as a carbon-carbon composite, which is a combination of graphite fibers and a carbonized resin matrix. Graphite is used in such applications as rocket nozzles or advanced turbine components. However, its disadvantage is a lack of oxidation resistance above 500 °C (930 °F). By using CVD to coat the graphite with silicon carbide, it is possible to provide oxidation protection up to 1650 °C (3000 °F). Silicon carbide oxidizes to form a silicon oxide surface that will prevent further oxidation. Advanced studies to further increase the oxidation temperature of graphite are in progress using such oxides as hafnia. Aluminum Coating of Steel. The coatings of steel (and other metals) with aluminum by ion plating is an industrial process with excellent potential for large-scale production. It replaces hot dipping of aluminum on steel, which has been used for many years. During immersion in hot dipping, a brittle intermetallic compound layer of iron-aluminum is formed; this limits the formability of the steel. In addition, the coating contains iron impurities, which may be a severe shortcoming in highly corrosive environments (Ref 4). Electroplated cadmium is another commonly used system for the corrosion protection of steel and other metals. However, it can cause hydrogen embrittlement of high-strength steel and titanium. In addition, cadmium has a low melting point (320 °C, or 608 °F) and presents environmental problems because of its toxicity. A process has recently been developed for applying aluminum on metals by ion-vapor deposition, which has none of the restrictions of hot dipping or cadmium electroplating. This process falls under military specification MIL-C 83488 (Ref 5). Vapor deposition of aluminum does not produce hydrogen; therefore, it does not cause embrittlement. Vapordeposited aluminum protects against stress-corrosion cracking (SCC) and has a temperature limit of 495 °C (925 °F). It is less expensive than most other barrier coatings. Applications for vapor-deposited aluminum in the aerospace industry include fuel and pneumatic line fittings (in place of anodizing), electrical connectors to replace cadmium, powder-metallurgy parts, fasteners, electrical black boxes, corrosion and SCC protection of depleted uranium and titanium alloys, and electromagnetic interference compatibility (Ref 6). One of the first industrial applications of the IBAD process was the coating of stainless steel electric razor screens with TiN for aqueous corrosion protection (Ref 7). Alloy Coatings for Gas Turbines. Gas turbines are used as aircraft turbines using highly refined fuels with operating temperatures between 900 and 1150 °C (1650 and 2100 °F) or higher; marine turbines using moderately refined fuels, which operate at 700 to 870 °C (1290 to 1600 °F) and are exposed to salt corrosion; and industrial turbines that may operate at 760 to 925 °C (1400 to 1700 °F) and use moderately refined fuels,

often with high sulfur contents. Industrial turbines may use a variety of fuels and operate in marine, arctic, or more benign environments. Along with the need for high-temperature mechanical strength, materials in these applications must resist oxidation and high-temperature corrosion. These problems are becoming more acute with the trend toward higher operating temperatures for better performance and improved fuel economy. Coatings were initially used in the 1960s to protect turbine hot section components from oxidation and hightemperature corrosion. Physical vapor deposition coating by electron-beam evaporation and sputtering of chromides and aluminides were used as temperature requirements increased, and these are still used in a majority of applications 30 to 40 years after their introduction. Overlay coatings of chromium, aluminum, and yttrium with cobalt, nickel, or iron are known generically as MCrAlY and are successfully used on aircraft turbines for resistance to hot corrosion. They are applied either by electron-beam PVD or plasma thermal spray. In the 1980s thermal barrier coatings (TBCs) were introduced for the purpose of reducing the operating temperature beneath the coating. Recent developments using platinum-modified aluminides increase the oxidation resistance of these thermal barriers (Ref 8). Marine gas turbines are subject to severe sulfidation, oxidation, and hot corrosion, especially in the presence of sodium sulfate (Na2SO4). Electron-beam-evaporated CoCrAlY coatings have considerably increased the service lives of first-stage blades. Further improvements have been achieved by using high-chromium low- yttrium alloys. Industrial turbines may not require the degree of corrosion protection of other applications unless high-sulfur fuels are used or the atmosphere is aggressive. In such cases, the hot-corrosion problem can be solved by using thin-film CoCrAlY coatings and a very large (200 kW) electron-beam coater capable of handling industrial turbine parts weighing up to 40 kg (90 lb). Now TBC coatings are used for industrial turbines in relatively lower temperature components. Failures of turbine blades due to corrosion-assisted fatigue can have catastrophic results even in the compressor portion of the turbine, and these can be protected with PVD coatings and IBED coating (Ref 9).

References cited in this section 4. F. Dunbar, “Aluminum Coated Steels: Past, Present and Future,” Paper 8201-037, presented at the ASM Metals Congress, Oct 1982 (St. Louis, MO), American Society for Metals 5. “Coating, Aluminum, High Purity,” MIL- DTL-83488D, April 1999 6. D. Muehlberger and J. Reilly, “Improved Equipment, Productivity Increase, Applications for Ion Vapor Deposition of Aluminum,” Technical Report 83-0691, Society of Automotive Engineers, 1985 7. G.K. Hubler and J.K. Hirvonen, Ion-Beam- Assisted Deposition, Surface Engineering, Vol 5, ASM Handbook, ASM International, 1994, p 599 8. M. Thomas, Materials Development in Aero Gas Turbines, Parsons 2000 Advanced Materials for 21st Century Turbines and Power Plant, A. Strang, W.M. Banks, R.D. Conoy, G.M. McMolvin, J.C. Neal, and S. Simpson, Ed., The University Press, Cambridge, United Kingdom, 2000, p 710 9. S. Chakravarty, P. Li, V. Tirabasso, and P.C Patnaik, Improvement of Fatigue and Corrosion Resistance of Compressor Rotor Blades of an Industrial Gas Turbine Engine, Advanced Materials and Coatings for Combustion Turbines, V.P. Swaminathan and N.S. Cheruvu, Ed., ASM International, 1994, p 135


Acknowledgment This article has been adapted from H.O. Pierson, CVD/PVD Coatings, Corrosion, Vol 13, ASM Handbook (formerly 9th ed. Metals Handbook), ASM International, 1987, p 456–458.


References 1. J.R. Davis, Ed., Surface Engineering for Corrosion and Wear Resistance, ASM International, 2001, p 175 2. G.K. Hubler and J.K. Hirvonen, Ion-Beam- Assisted Deposition, Surface Engineering, Vol 5, ASM Handbook, ASM International, 1994, p 593 3. A.H. Deutchman and R.J. Partyka, Comparison of the Properties of PVD and IBED Hardcoats (TiN and Cr2N), First ASM International Surface Engineering Congress and 13th IFHTSE Congress, Oct 2002 (Columbus, OH), ASM International 2003, p 138 4. F. Dunbar, “Aluminum Coated Steels: Past, Present and Future,” Paper 8201-037, presented at the ASM Metals Congress, Oct 1982 (St. Louis, MO), American Society for Metals 5. “Coating, Aluminum, High Purity,” MIL- DTL-83488D, April 1999 6. D. Muehlberger and J. Reilly, “Improved Equipment, Productivity Increase, Applications for Ion Vapor Deposition of Aluminum,” Technical Report 83-0691, Society of Automotive Engineers, 1985 7. G.K. Hubler and J.K. Hirvonen, Ion-Beam- Assisted Deposition, Surface Engineering, Vol 5, ASM Handbook, ASM International, 1994, p 599 8. M. Thomas, Materials Development in Aero Gas Turbines, Parsons 2000 Advanced Materials for 21st Century Turbines and Power Plant, A. Strang, W.M. Banks, R.D. Conoy, G.M. McMolvin, J.C. Neal, and S. Simpson, Ed., The University Press, Cambridge, United Kingdom, 2000, p 710 9. S. Chakravarty, P. Li, V. Tirabasso, and P.C Patnaik, Improvement of Fatigue and Corrosion Resistance of Compressor Rotor Blades of an Industrial Gas Turbine Engine, Advanced Materials and Coatings for Combustion Turbines, V.P. Swaminathan and N.S. Cheruvu, Ed., ASM International, 1994, p 135


Selected References Articles in Surface Engineering, Volume 5 of the ASM Handbook, ASM International, 1994: • • • • • • • • • •

J.K. Hirvonen and B.D. Sartwell, Ion Implantation, p 605–610 J.S. Horwitz, Pulsed-Laser Deposition, p 621–626 D.M. Mattox, Growth and Growth-Related Properties of Films Formed by Physical Vapor Deposition, p 538–555 D.M. Mattox, Ion Plating, p 582–592 D.M. Mattox, Vacuum Deposition, Reactive Evaporation, and Gas Evaporation, p 556–572 H.O. Pierson, Chemical Vapor Deposition of Nonsemiconductor Materials, p 510–516 M. Razeghi, Chemical Vapor Deposition of Semiconductor Materials, p 517–531 S.L. Rohde, Sputter Deposition, p 573–581 D.M. Sanders, J.W. Glaser, and S. Falabella, Arc Deposition, p 602–604 P.K. Tedrow and R. Reif, Plasma-Enhanced Chemical Vapor Deposition, p 532–537

Other reference source: • • • • • • •

G. Decher and J.B. Schlenoff, Ed., Multilayer Thin Films: Sequential Assembly of Nanocomposite Materials, Wiley-VCH, 2003 M.L. Hitchman and K.F. Jensen, Ed., Chemical Vapor Deposition, Academic Press, 1993 W. Menz and J. Mohr, Microsystem Technology, Wiley-CVH, 2000 J.-H. Park and T.S. Sudarshan, Ed., Chemical Vapor Deposition, ASM International, 2001 H.O. Pierson, Handbook of Chemical Vapor Deposition, Noyes Publication, 1992 R.V. Stuart, Vacuum Technology, Thin Films, and Sputtering: An Introduction, Academic Press, 1983 V.G. Syrkin, The CVD Method: Chemical Vapor-Phase Metallization, 2000

V.A. Ravi, Pack Cementation Coatings, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 763–771

Pack Cementation Coatings Vilupanur A. Ravi, California State Polytechnic University, Pomona

Introduction METALLIC MATERIALS used in harsh environmental conditions undergo degradation in various ways. Several forms of corrosion, that is, uniform, galvanic, pitting, and crevice corrosion (Ref 1), have been observed, reported, and analyzed in aqueous environments. Similarly, at high temperatures, alloys can degrade in different ways: by oxidation (Ref 2), sulfidation (Ref 3), halogen attack (Ref 4), and hot corrosion (Ref 5). Two general approaches are used to mitigate the corrosive effects of high-temperature environments:

• •

Alloy design: The alloy composition and/or structure are altered in specific ways to address the specific environment. Surface modification: Coatings are applied where the surface structure and/or composition is engineered appropriately while retaining the same substrate.

Coatings containing aluminum, chromium, and silicon are effective against oxidation and hot corrosion by forming protective Al2O3, Cr2O3, and SiO2 scales, respectively. Some of the key characteristics of a successful coating are: • • •

The ability to resist damage in the desired service conditions (corrosion, erosion, erosion- corrosion, stress-related corrosion problems, etc.) The integrity of the coating-substrate interface (an ideal interface would be continuous and free from interfacial stresses) The adherence of the coating to the substrate under mechanical and thermal cycles

Several coatings and associated processes are available to protect metallic substrates from a range of corrosive environments. Table 1 (Ref 6) indicates one method of classification of coating processes. Although not exhaustive, the table illustrates an important distinction between two major classes of coatings: overlay and diffusion (surface modified). In the former, the coatings are applied with minimal change to the substrate. In the latter, however, the coating constituents diffuse into the substrate, resulting in chemical and microstructural changes to the substrate surface. Table 1 Classification of coating processes Category Overlay coatings

Description Processes Deposition with minimal interdiffusion of • Hot spraying (flame spraying, plasma coating with substrate spraying, detonation coatings) • Chemical vapor deposition

Surface modified coatings

• Physical vapor deposition (thermal evaporation, ion sputtering, ion plating) Diffusion-induced modification of chemical • Pack cementation composition and/or microstructure of substrate • Slurry cementation • Metalliding • Hot dipping • Surface alloying (ion implantation, laser surface alloying)

Source: Ref 6 Each coating and its associated processes has features that make it suitable for certain applications and unsuitable for others. The focus in this article is pack-cementation coatings and, in particular, halide-activated pack cementation coatings on nickel alloys, although the principles are generally applicable to Fe- and Co-base alloys.

References cited in this section 1. M.G. Fontana and N.D. Greene, Corrosion Engineering, 2nd ed., McGraw-Hill International Book Company, 1978, p 28 2. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science Publishers, Ltd., 1988, p 324

3. N.S. Bornstein, JOM, J. Min., Metals, Mater. Soc., Vol 48, 1996, p 37 4. P. Elliott, C.J. Tyreman, and R. Prescott, J. Met., Vol 37, 1985, p 20 5. R.A. Rapp, Corrosion, Vol 42, 1986, p 568 6. P.G. Capelli, Coating Processes, High Temperature Alloys for Gas Turbines, D. Coutsouradis, P. Felix, H. Fischmeister, L. Habraken, Y. Lindblom, and M.O. Speidel, Ed., Applied Science Publishers, London, 1978, p 177

V.A. Ravi, Pack Cementation Coatings, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 763–771 Pack Cementation Coatings Vilupanur A. Ravi, California State Polytechnic University, Pomona

Halide-Activated Pack Cementation Halide-activated pack cementation (Ref 7) is a coating process in which the elements to be deposited are converted into halide vapor species. These halide vapors then diffuse through a porous pack to the surface of the substrate to be coated, where they undergo disproportionation reactions resulting in the deposition of the desired elemental species. The coating material then diffuses into the substrate with subsequent modification of the substrate microstructure. A typical pack (Fig. 1) (Ref 8) consists of: • • • •

Filler Master alloy Activator salt Substrate to be coated

Powders of the filler, master alloy, and activator are blended together and placed in a container with the substrate buried in it.

Fig. 1 Typical pack of constituents for a laboratory scale cementation pack A filler, such as Al2O3, allows the active pack constituents (the master alloy and activator salt) to be well distributed. This allows the halide vapors to easily reach the substrate surface, resulting in an overall uniform coating. The activator salt, a halide compound such as NaCl, CrCl2, NH4Cl, or AlF3, reacts with the master alloy to produce gaseous halide compounds of the element(s) to be deposited. The master alloy, such as Fe-Al or Ni-Al, contains the element(s) to be deposited. However, the elemental forms of Al, Cr, or Si can also be used. The alloy compositions are customized to achieve the desired coating composition Substrates are typically nickel- or iron-base alloys. Their forms can have a wide range of geometries, for example, flat plates, pipes, and turbine blades. It is important that the substrates are not positioned too close together in the pack, to avoid nonuniformity in the coating thickness.

References cited in this section 7. L.L. Seigle, Surface Engineering, R. Kossowsky and S.C. Singhal, Ed., Martinus Nijhoff Publishers, Pittsburgh, 1984, p 345 8. V.A. Ravi, Ph.D. thesis, The Ohio State University, 1988


Types of Pack Cementation Processes Pack aluminizing, chromizing, and siliconizing are some distinct classes of pack cementation coating processes. These processes are based on coating the substrate with a single element, that is, aluminum, chromium, or silicon. It is also possible to achieve simultaneous deposition of more than one element.

Pack Aluminizing The original process, attributed to Van Aller (Ref 9), used a pack consisting of aluminum, ammonium chloride, and graphite powders placed in a drum along with iron or copper substrates. The coating process was carried out around 450 °C (840 °F). Gilson (Ref 10) replaced the graphite powder with aluminum oxide. These early processes remain essentially the same today. The current versions, however, involve a range of master alloys and activators. Other researchers (Ref 11, 12, 13) have classified aluminizing packs, for example, as “high-” or “low-” activity packs. In high-activity packs, aluminum diffuses inward into the substrate (Fig. 2) and forms a solid solution intermetallic such as NiAl with the nickel in the substrate. In low-activity aluminum packs, on the other hand, the outward diffusion of nickel from the substrate dominates. Figure 3 (Ref 11) depicts the microstructure of a nickel-base superalloy aluminized in a low-activity aluminum pack.

Fig. 2 Archtypical microstructure of a cross section of a nickel-base superalloy aluminized in a highactivity aluminum pack followed by heat treatment for 4 h at 1080 °C (1975 °F). See source Ref 11 for exact size

Fig. 3 Archtypical microstructure of the cross section of a nickel-base superalloy aluminized in a lowactivity aluminum pack. See source Ref 11 for exact size Figures 4 and 5 (Ref 12) are schematics of coating formation in low- and high-activity Al packs as discussed in Ref 12. Figure 4(a) shows markers placed on the surface of pure nickel. When exposed to a low-activity aluminum pack, the outward diffusion of nickel is much faster than the inward diffusion of aluminum. Thus the coating grows outward, as indicated by intermediate positions e1 and e2. The net outward diffusion of nickel results in the formation of Kirkendall pores close to the original interface. Figure 4(b) (Ref 12) represents the situation for a nickel-base superalloy in which the outward diffusion of nickel leads to corresponding nickeldepleted areas within the alloy. Thus when a layer (position e1) forms by outward nickel diffusion, a corresponding layer of nickel depletion (position

) forms within the alloy. Likewise, e2 corresponds to

, e3

corresponds to , and so forth. The loss of nickel leads to the formation of intermetallics like NiAl and is accompanied by precipitates like TiC, which are not soluble in the intermetallic matrix. This is further illustrated in Fig. 4(c).

Fig. 4 Schematic showing progressive stages of aluminization in a low-activity aluminum pack. (a) Pure nickel (e1 forms first, followed by e2, e3, etc.). (b) Nickel-base superalloy (e1 and e2 and

, etc.). (c) Structure of coated nickel-base superalloy. Source: Ref 12

form first, followed by

Fig. 5 Schematic showing progressive stages of aluminization in a high-activity aluminum pack. (a) Pure nickel (e1 forms first, followed by e2, e3, etc.). (b) Nickel-base superalloy e1, followed by e2, etc. Source: Ref 12 In high-activity aluminum packs, inward diffusion of aluminum dominates as illustrated in Fig. 5 (Ref 12). Figure 5(a) shows the position of the markers originally placed on the free surface of nickel, moving to e1, e2, e3, and so forth, as a function of time; that is, the markers remain on the outer surface. The inward diffusion is accompanied by the formation of an intermetallic like Ni 2Al3. Figure 5(b) shows the progress of aluminization in a nickel-base superalloy. The markers stay on the outer surface, indicative of the inward diffusion of aluminum. The formation of aluminum-rich Ni2Al3 leads to the precipitation of chromium-rich precipitates in the matrix alloy. Figure 6 (Ref 12) shows the effect of annealing on aluminized nickel (Fig. 6a) and nickel- base superalloys (Fig. 6b). In both cases, the Ni2Al3 layer transforms to NiAl. In the case of pure nickel, the conversion of Ni2Al3 to NiAl is fairly straightforward. However, in the case of superalloys, this transformation is characterized by precipitates in the outer zone, a central precipitate-free zone, and an inner zone with precipitates similar to those for low-activity aluminum packs.

Fig. 6 Schematic showing progressive stages of aluminization in a high-activity aluminum pack followed by diffusion annealing. (a) Pure nickel. (b) Nickel-base superalloy. Source: Ref 12 Others (Ref 14) have shown that the structures in Fig. 2 and 3 are limiting cases with a range of intermediate structures existing due to the variation of the interdiffusion coefficient in the β- NiAl phase of the aluminumnickel system. Thermodynamics of Pack Aluminizing. If a pure nickel substrate is aluminized in a pack containing pure aluminum (thermodynamic activity = 1) as the source, and NaCl is the activator, the following equilibria pertain: Al(l) + NaCl(l)

AlCl(v) + Na(v)

(Eq 1)

AlCl2(v) and AlCl3(v) are also formed in the pack. The most important reaction occurring on the substrate surface is the disproportionation of the metal halide vapors AlCl or AlCl2: (Eq 2) where Ni indicates Al is in solid solution in the Ni substrate. Direct transfer of aluminum by forming Al(v) occurs only to a very small extent. Displacement reactions with the substrate such as: Ni + 2AlCl(v)

NiCl2(v) + 2AlNi

(Eq 3)

do not occur because the standard Gibbs energies of formation of the aluminum halides are significantly more negative than those for the nickel halides (Ref 15). Thus, if pure aluminum is used as the source for aluminizing pure nickel, the NiCl and NiCl2 partial pressures are significantly lower than the partial pressures of AlCl, AlCl2, and AlCl3 (per mole of Cl2) so that nickel does not participate in interchange reactions with aluminum chloride. The chemical reactions in the pack are only one of a series of steps in pack cementation processes. Walsh (Ref 16) has broken down the pack cementation process into the following steps: 1. Reactions in the pack 2. Gas diffusion through the pack

3. Reaction on the substrate surface, which can involve disproportionation of halide vapors, precipitation of activator salt, and/or reactions with the environment and substrate Therefore, in addition to understanding the thermodynamics, it is also necessary to analyze the kinetics to gain a full appreciation of the process. Kinetics of Pack Aluminizing. Several models for the aluminizing of nickel and iron-base alloys have been proposed (Ref 17, 18, 19, 20). All of them are based on the existence of a depletion zone. A kinetic model of pack aluminizing based on the assumption of local thermodynamic equilibrium at both the pack-gas and the substrate-gas interfaces was proposed (Ref 17). The authors also assumed that the transport of gases through the pack was purely diffusional. Instantaneous fluxes of the species to and from the substrate were calculated under conditions of 1 atm total pressure and an activity of aluminum equal to 0.01 at the substrate surface. Figure 7 (Ref 17) shows the model for NaX and NH4X activated packs where X = F, Cl, Br, or I. The diffusion direction for each species in the depletion zone is obtained by computing the instantaneous fluxes. Then, depending on the temperature and the choice of X, certain pressure gradients (e.g., , ) will be opposite to those shown schematically. Figure 7(a) is a simplified model in which the formation of liquid NaX is neglected. Figure 7(b) is more complete since the liquid NaX film on the substrate is taken into account while computing fluxes. Figure 7(c) is a complete model for NH4X-activated packs. The direction of the diffusion of AlX2, AlX3 and Al12X6 depends on X. In NH4F activated packs, AlF and AlF2 diffuse inward while AlF3 and Al2F6 diffuse outward. In NH4Cl, NH4B4 and NH4I activated packs, AlX alone diffuses inward. In all cases, aluminum deposition in the NH4X and NaX-activated packs is primarily done by AlX. The NaX(l) condensation on the substrate surface aids in aluminum deposition since there is no requirement for halogen removal by the diffusion of AlX2 and AlX3 away from the substrate surface.

Fig. 7 The Levine-Caves model for pack aluminization using pure aluminum as the source and NaX in (a) and (b) or NH4X in (c) as activators (X is Cl, Br, or I). Part 7(a) is a simplified model and (b) and (c) are more complete models. Source: Ref 17 The model described in Ref 17 was applied to the aluminizing of the IN 100 superalloy. Loss of alloying elements during the coating treatment, especially chromium, had to be considered. The nickel and chromium can be removed from the surface either in the metal vapor form or by forming volatile halides. The loss of nickel and chromium only by forming a metal vapor was considered. The nickel flux leaving the substrate is negligible while the chromium flux is about 1% of the aluminum weight gain. The loss of nickel by forming a halide would also be negligible, but the loss of chromium by forming chromium halides could make the chromium loss somewhat higher than the 1% used in Ref 17. However, this model is comprehensive and provides an adequate description of the pack aluminizing of superalloys. Other authors (Ref 18) point out that in a NaX or AlF3 activated pack, condensation of activators such as NaX(l) or AlF3(s) on the substrate result in a net flux of activator from the pack to the specimen. Therefore, in addition to an aluminum-depleted zone, an activator-depleted zone forms adjacent to the substrate. If the amounts of activator and aluminum were stoichiometrically comparable, the activator- and aluminum-depletion zones would coincide. However, the amount of activator used in pack aluminizing is much less than the amount of aluminum, and the activator-depleted zone is thus wider than the aluminum-depleted zone (Fig. 8). These researchers have shown that it is possible to maintain a minor difference in total pressure between the pack-gas interface and the substrate- gas interface.

Fig. 8 The Kandasamy-Seigle-Pennisi model with mixed aluminum transport through activator and activator + Al-depleted zones. Source: Ref 18 Another pair of researchers (Ref 19) modified the two models described earlier by including bulk flow of the vapors as an additional feature. This flow occurs due to the difference in the number of moles of gas involved in reactions at the interface. However, no difference in total pressure was allowed, as in the model described in Ref 18. Other researchers (Ref 20) have developed a more complete model taking into account the total pressure differential (Ref 18 model) as well as the viscous flow of the halide vapors through the porous pack (Ref 19 model) (Fig. 9). Aluminum transport rate constants and partial pressure values for NH4Cl and AlF3 activated packs have been calculated. The values obtained fall between those obtained by the models described in Ref 19 and 17, which represent limiting cases for packs with zero resistance to viscous flow (large particle size of the pack powders) and infinite resistance to viscous flow (small particle sizes).

Fig. 9 The Wang-Seigle model for aluminum transport incorporating diffusional flux and viscous flow in the depleted zone. Source: Ref 20 The aluminizing of iron through a fixed porous barrier (ceramic felt wrapping) was studied (Ref 21), and a model based on the following assumptions was proposed: (a) local equilibrium at the pack-gas and substrategas interfaces, and (b) rate control by diffusion of gases through a porous barrier surrounding the substrate in series with solid-state diffusion and condensation of activators such as NaCl(l) and AlF3(s) on the specimen surface. Figures 10 and 11 (Ref 21) indicate the direction of fluxes of the species involved. The authors of Ref 21 did not consider a total pressure difference between the pack-gas interface and the substrate-gas interface but did consider viscous flow. Good agreement between theoretical and experimental values for rates of aluminum gain was obtained for NaCl-activated packs.

Fig. 10 The Kung-Rapp model for transport in a NaCl-activated pack with an Fe-Al alloy source and with the depletion zone corresponding to the thickness of the alumina wrapping. Source: Ref 21

Fig. 11 The Kung-Rapp model for transport in an AlF3-activated pack with an Fe-Al alloy source and with the depletion zone corresponding to the thickness of the alumina wrapping. Source: Ref 21 Others (Ref 22) have determined the surface aluminum concentration based on the formation of a liquid aluminum-enriched layer of variable aluminum composition on an iron substrate (Fig. 12). The two novel features of this model are: (a) the formation of a thin liquid metal layer of almost constant thickness on the substrate (allowing an aluminum flux balance to be set up), and (b) the concentration of aluminum at the transition boundary layer, M, given by the solidus composition of the iron-aluminum equilibrium diagram. The model prediction in Ref 22 of the surface aluminum concentration agrees reasonably well with experimentally obtained values.

Fig. 12 The Akueze-Stringer model for the aluminization of pure iron. Source: Ref 22

Pack Chromizing Kelley (Ref 23) is generally acknowledged as the inventor of pack chromizing. The enhanced resistance to aqueous corrosion of these coatings was initially used to improve the surface stability of steam turbine blades. Later, chromium coatings were used to protect gas turbine blades and vanes against hot corrosion (Ref 13). A typical chromium coating on a nickel-base superalloy, coated in a pack containing chromium, ammonium chloride, and alumina, is a γ- Ni-Cr solid solution with a discontinuous α-Cr outer layer. This outer layer contains Al2O3 particles that are attributed in Ref 11 to internal oxidation of aluminum in the substrate. Ref 11 further points out that other investigators (Ref 24, 25) have prevented this by adding aluminum to the chromium source in the pack. In Ref 24, a mixture of nickel, cobalt, chromium, and aluminum powders was preheated in a pack containing NH4Cl, MgCl2, and Al2O3 to approximately 1040 °C (1900 °F) and held it there for 20 h. The pack was then cooled and substrates introduced before reheating to 1040 °C in a hydrogen atmosphere. In Ref 25, the coating treatment was performed in air with chromium and aluminum present in the pack. The aluminum in the pack functions as an oxygen getter and permits the chromium coating to be applied without internal oxidation. Literature on the structure and properties of chromium coatings for nickel-base superalloys is sparse. References 26 and 27 present structural and kinetic data for chromide coatings on steel, but these are expected to be different for nickel- base alloys. Thermodynamics of pack chromizing. Hoar and Croom (Ref 28) have suggested that incorporation of chromium into substrates was achieved by one or more of the following reactions:displacement reactions with the substrate

(Eq 4) where

is in solid solution and reduction of halide vapor (in a reducing hydrogen atmosphere): (Eq 5)

Thermal decomposition of halide vapor is represented as: (Eq 6) Displacement reactions are important only if the substrate element forms volatile halide vapors of appreciable partial pressure (e.g., iron). Displacement reactions with the substrate such as Ni: (Eq 7) are not expected to occur because the standard Gibbs energies of formation of the chromium halides are more negative than those for the nickel halides (Ref 15, 29). Thus, if pure chromium is used as the source for chromizing pure nickel, the NiCl and NiCl2 partial pressures are lower than the partial pressures of CrCl2 and CrCl3 (per mole of Cl2) so that nickel is not expected to participate in interchange reactions with chromium chloride. In Ref 28, it is pointed out that since displacement reactions are insignificant for nickel substrates, a hydrogen atmosphere is essential for chromizing nickel. However, it is possible that a reaction of the type: (Eq 8) could provide for chromium transport. Therefore, it should be possible to chromize nickel in an inert atmosphere. In Ref 30, pure nickel in NH4Cl-activated packs containing pure chromium was chromized. The investigator correctly points out from free energy considerations that displacement reactions of the type shown in Eq 7 are not responsible for the coating process. The most important coating reaction was attributed to: (Eq 9) Once again, the possibility of disproportionation of CrCl2 was not considered. Another researcher (Ref 14) points out that the presence of HCl in NH4Cl-activated packs results in a side reaction: (Eq 10) with the inert filler. The H2O(v) can undergo a further reaction: (Eq 11) with condensed CrCl2. Kinetics of Chromizing. Mazille (Ref 30) investigated the kinetics of chromizing pure Ni in an ammonium chloride pack containing pure chromium as the source. He obtained an outer α-Cr layer after chromizing for 16 h at 950 °C (1740 °F). This outer layer is an α-Cr region with an inner single-phase, Ni-rich γ-solid solution. It was estimated that the chromium content of the outer layer exceeded 90% by weight Cr. From the activitycomposition plot for chromium in a nickel-chromium alloy, shown in Fig. 13 (Ref 31) this composition is in the two-phase α + γ region and the activity of chromium is approximately 0.96.

Fig. 13 The activity of chromium in solid nickel-chromium alloys (referred to as solid pure Cr) for various temperatures. Source: Ref 31 From Fig. 14(a), it was concluded that the chromizing of nickel is controlled by the slow solid state diffusion into austenitic nickel and not by transport in the gas or reactions at the substrate surface. The change of slope for the 950 °C (1740 °F) curve is attributed to the slower formation of the external α-phase when compared to the 1000 or 1050 °C (1830 or 1920 °F) treatments. Once α forms, diffusion occurs more rapidly since it is a body-centered cubic structure.

Fig. 14 Kinetic data for the chromizing of pure Ni. (a) The specific weight gain as a function of time at different temperatures. (b) Variation of thickness of the inner layer with time at different temperatures. Source: Ref 30 Figure 14(b) shows the growth rate of the γ- phase. While a parabolic relation is observed at 1000 and 1050 °C (1830 and 1920 °F), the shape of the 950 °C (1740 °F) curve is different. This is attributed to the approximately equal diffusion rates through the γ-α and γ-pure nickel phase boundaries. This suggests that the inner γ-layer remains approximately of constant thickness once the slower growing outer α-layer forms. Information on the effects of coating variables such as the type of halide activators and fillers on coating characteristics such as microstructure and composition for the chromizing of ferrous alloys can be found in Ref 32.

Simultaneous Chromizing-Aluminizing

Initial studies in the area of simultaneous chromizing and aluminizing were directed toward a two-step chromizing-aluminizing process. Chromium coatings were initially obtained in a pack containing pure chromium, and the chromized coupons subsequently were aluminized in low- or high-aluminum activity packs (Ref 33). Galmiche (Ref 34, 35) proposed a technique for simultaneous chromizing-aluminizing of nickel-, iron-, and cobalt-base alloys. The process, called chromaluminization, consists of heating the substrates in a pack containing ultrafine chromium powder (40 wt%), very fine aluminum powder (10 wt%), fine-calcined alumina (49.2 wt%), and ammonium chloride (0.8 wt%) to temperatures in the range 750 to 1200 °C (1380 to 2190 °F). Figure 15 (Ref 34) shows the composition profile of a chromaluminized coating on a nickel-base superalloy. Galmiche explained that chromium is preferentially supplied during the first “phase” of the “chromaluminizing” treatment, while aluminum is preferentially supplied during the second “phase.” This is somewhat unexpected because the partial pressures of aluminum halides are significantly higher than those of the chromium halides, and a pack containing a mixture of pure chromium and pure aluminum would thus be expected to aluminize the substrate.

Fig. 15 Concentration profiles of chromium and aluminum in the chromaluminized regions of a nickelbase superalloy. Source: Ref 34 Galmiche and Hivert's patent (Ref 35) reveals that pure Cr (40 wt%) and pure Al (40 wt%) were preheated in a pack containing alumina and ammonium chloride for 1 h in the temperature range 1000 to 1100 °C (1830 to 1210 °F). Under these conditions, aluminum chloride vapor pressures are much higher than the chromium chloride vapor pressures, and it follows that aluminum ought to be deposited on the chromium particles. It is also indicated (Ref 35) that additions of 0.5 to 1.0 wt% AlF3 to the pack will be advantageous to the coating process. It is known that the AlF3 is a strong aluminizing agent. It is thus possible that during pretreatment, aluminum is deposited on the chromium particles by AlF3 and AlCl3 (NH4Cl-assisted) vapors. During the coating process, a significant amount of AlCl3 vapors are expected to be generated. Furthermore, in the proposed mechanism, the aluminum-coated chromium particles are expected to generate only marginal amounts of chromium-containing vapors. Conditions in the pack are thus expected to be conducive for aluminization, while conserving the chromium content of the alloy. Based on this model, the sharp increase in chromium content in the inner diffusion zone (Fig. 15) can be attributed to outward nickel diffusion and consequent nickel depletion in this region. Other authors (Ref 16, 36) have pointed out that it is not possible to codeposit aluminum and chromium in a one-step process if pure aluminum and pure chromium are used as the source. Others (Ref 29) have shown that at temperatures in the 1100 to 1500 K (1520 to 2240 °F) range, where most coating operations are carried out, the vapor pressures of the AlCl species (mostly responsible for aluminizing) are orders of magnitude higher than the CrCl2 vapor pressures (mostly responsible for chromizing) in chloride- activated packs using pure aluminum and pure chromium as the sources. They obtained even greater differences in aluminum- and chromium- fluoride vapor pressures in fluoride-activated packs, concluding that at unit thermodynamic activity of aluminum, codeposition of aluminum and chromium is not possible. A clear perspective on the difficulties of codeposition of aluminum and chromium is obtained by examination of the thermodynamic activities of aluminum and chromium in binary aluminum-chromium alloys. Others (Ref 37) have determined the activities of aluminum and chromium in binary aluminumchromium alloys (Ref 38) (Fig. 16). The maximum solubility of aluminum in chromium is 43 a/o aluminum; for aluminum compositions lower than this, the activity of aluminum is much lower than that of chromium in chromium-rich, chromium-aluminum alloys. This highly negative deviation from thermodynamically ideal behavior is

demonstrated by the low aluminum activity (0.001) at 9 at.% Al. Therefore, adjusting the composition of a chromium-aluminum master alloy is one of the keys to achieving codeposition. In addition, the choice of the activator is also important.

Fig. 16 Activities of aluminum and chromium as a function of the aluminum content of aluminumchromium alloys at 1273 K (1830 °F). Source: Ref 36, based on data from Ref 37 Still another investigator (Ref 39) has shown experimentally that by choosing a special pack composition, simultaneous deposition of chromium and aluminum into a nickel-base alloy at 1100 °C (2010 °F) is possible. The following reaction is proposed as the basis for codeposition: AlCl2(v) + Cr

2CrCl2(v) + Al

(Eq 12)

From Eq 12, it follows that the partial pressure of CrCl2 is given by: (Eq 13) where Kp is the equilibrium constant, aCr and aAl are the activities of chromium and aluminum, respectively, and and are the partial pressures of the chromium and aluminum chlorides. By reducing the thermodynamic activity of aluminum (aAl) in the pack, one should expect a reduction in the partial pressure of the aluminum halides, thus leading to codeposition of aluminum and chromium. Ravi et al. ( 8, 40, 41, 42) have shown that chloride activators and chromium-rich chromium-aluminum master alloys result in chromium-rich chromium-aluminum codeposited coatings on nickel and nickel-base superalloy substrates. They also show that the use of fluoride activators favors aluminum deposition, even with the use of chromium-rich chromium-aluminum master alloys. Detailed thermodynamic calculations and microstructural/compositional analysis are provided. Other researchers (Ref 43) have identified process conditions (chromium-aluminum master alloy composition, NaCl/NH4Cl ratios) for achieving simultaneous deposition of chromium and aluminum at temperatures on the order of 1100 °C (2010 °F) onto a nickel-base superalloy. They established that codeposition of chromium and aluminum can be achieved with chromium-rich chromium-aluminum master alloys and mixed activator (NaCl + NH4Cl) packs. Also studied were the kinetics and nonequilibrium effects in the coating process (Ref 44, 45). One of the key findings was that the master alloy composition and activator-mixture content change during the coating process. Based on these findings, the authors have proposed a time-dependent mechanism for chromium-aluminum codeposition. Other authors (Ref 46) have developed a single-step process for chromium/rare-earth (RE) modified and REdoped aluminide coatings by using appropriate chloride activators and chromium-aluminum master alloys. A model for coating formation based on their analyses of the microstructures, phase equilibria, and kinetic data

was proposed. The results of experiments aimed at reducing pack entrapment (Al2O3 and RE oxide particles) in the coatings have been discussed. Results from environmental testing of these coated coupons under different conditions, that is, isothermal oxidation at 1100 °C (2010 °F) in air, cyclic oxidation at 1100 °C in static air, and type 1 hot-corrosion behavior at 900 °C (1650 °F) in 0.1% SO2/O2 gas mixtures have been discussed (Ref 47). In general, the Cr/RE- doped coatings resulted in improved performance relative to RE-free coatings. Another pair of researchers (Ref 48) coated a 2.25 Cr-1Mo low alloy steel with RE modified chromiumaluminide coating. They report improved cyclic oxidation performance of these steels in air at 800 °C (1470 °F) relative to the uncoated alloy. Kim et al. (Ref 49) described a new pack process for the codeposition of chromium and aluminum. In this process, low-alloy steels were coated in a pack containing aluminum and Cr2O3 powders blended with an activator salt mixture of NaCl + NH4Cl. The authors propose that chromium halides are produced as a result of reactions between the aluminum or aluminum halides and the Cr2O3, leading to deposition of chromium along with the expected deposition of aluminum by aluminum halides created by reaction between the aluminum and the activator salts. The authors present microstructures and compositional profiles, although detailed thermodynamic evaluations have not been carried out, and have also extended the work by coating austenitic stainless steels (Ref 50).

Pack Siliconizing Another researcher (Ref 51) siliconized pure iron and steels in packs containing pure silicon and NH 4Cl as the activator. Gorbunov (Ref 52) has reported microstructural and kinetic data for siliconized iron and steel. The siliconizing was performed using pure silicon powder and gaseous chlorine. Recently, Hsu and Tsai (Ref 53) reported that siliconized AISI type 310 stainless steel (UNS S31000) has shown better resistance to hightemperature corrosion in air and in reducing gases in comparison with the base alloy. However, there are several problems with siliconized coatings. It has been pointed out (Ref 54) that silicon can cause embrittlement of the substrate. In addition, silicon can form low- melting oxide eutectics with iron and low-melting metallic eutectics with nickel substrates. Therefore, industry has not widely adopted siliconizing due to some of the aforementioned problems.

Simultaneous Chromizing-Siliconizing The codeposition of chromium and silicon onto iron-base alloys by halide-activated pack cementation has been reported (Ref 55). Different binary chromium-silicon master alloys, ranging from a minimum of 10 wt% Si to a maximum of 40 wt% Si, and different activators (NaCl, NaF, FeCl2, and CrCl2) were used in the study. Its authors performed a comprehensive computer-assisted thermodynamic analysis to obtain insights into pack chemistries. Using these computations as a guide, they selected appropriate pack chemistries and presented the microstructures of coated plain carbon, low-alloy, and austenitic steels. The authors report that selected coated coupons performed very well in cyclic oxidation and in oxidation-sulfidation tests. Germanium- and boron-doped silicide coatings on niobium- and titanium-base alloys have been studied in great detail (Ref 56, 57). The thermodynamics, kinetics of growth, and oxidation studies in these materials are reported. Cockeram and Rapp (Ref 58) also report on the thermodynamics, microstructural analysis, and kinetics of titanium silicide coatings formed on commercially pure titanium. Others (Ref 59) have studied the deposition of aluminide and silicide coatings on γ-TiAl.

Simultaneous Aluminizing-Siliconizing A pack cementation process for the codeposition of aluminum and silicon onto a low carbon steel has been developed (Ref 60). The pack contains aluminum and SiO2 powders and is activated by a NaCl+KBF4 mixture. Compositional profiles and microstructures of the coated coupons are reported. The authors propose that the following reaction is responsible for silicon deposition, although no detailed thermodynamic analysis is provided: 3SiO2 + 4Al

3Si + 2Al2O3

(Eq 14)

Xiang et al. (Ref 61) report that codeposition of aluminum and silicon onto γ-TiAl is possible in packs containing aluminum and silicon powders (in elemental form) and NH4Cl and AlCl3 activators. They have

shown microstructures that contain an outer silicide layer and an inner aluminide layer and propose a sequential mechanism, that is, aluminum deposition followed by silicon deposition, to explain the two-layer morphology. They also report that results of oxidation tests were favorable.

References cited in this section 8. V.A. Ravi, Ph.D. thesis, The Ohio State University, 1988 9. T. Van Aller, U.S. Patent 1,155,974, 1911 10. E.G. Gilson, U.S. Patent 1,091,057, 1914 11. G.W. Goward and D.H. Boone, Oxid. Metals, Vol 3, 1971, p 475 12. R. Pichoir, Influence of the Mode of Formation on the Oxidation and Corrosion Behavior of NiAl-Type Protective Coatings, Materials and Coatings to Resist High Temperature Corrosion, D.R. Holmes and A. Rahmel, Ed., Applied Science Publishers Ltd., London, 1978, p 271 13. G.W. Goward and L.W. Cannon, Trans. ASME, Vol 110, 1988, p 150 14. S. Shankar and L.L. Seigle, Met. Trans. A, Vol 9, 1978, p 1468 15. L.B. Pankratz, J.M. Stuve, and N.A. Gokcen, Thermodynamic Data for Mineral Technology, Bureau of Mines Bulletin 677, U.S. Government Printing Office, 1984 16. P.N. Walsh, Chemical Aspects of Pack Cementation, Proc. Fourth International Conf. on Chemical Vapor Deposition, G.F. Wakefield and J.M. Blocher, Ed., The Electrochemical Society, Pennington, N.J., 1973, p 147 17. S.R. Levine and R.M. Caves, J. Electrochem. Soc., Vol 121, 1974, p 1051 18. M. Kandasamy, L.L. Seigle, and F.J. Pennisi, Thin Solid Films, Vol 84, 1981, p 17 19. B. Nciri and J. Vandenbulcke, J. Less Common Met., Vol 95, 1983, p 191 20. T.H. Wang and L.L. Seigle, The Influence of Viscous Flow on the Kinetics of Gas Transport in Aluminizing Pack, High Temperature Protective Coatings, M. Khobaib and R.C. Krutenat, Ed., The Metallurgical Society, Warrendale, PA, 1987, p 39 21. S.C. Kung and R.A. Rapp, J. Electrochem. Soc., Vol 135, 1988, p 731 22. H.C. Akuezue and J. Stringer, J. Electrochem. Soc., Vol 135, 1988, p 1290 23. F.D. Kelley, Trans. Amer. Electrochem. Soc., Vol 43, 1923, p 351 24. A.L. Baldi, U.S. Patent 4,308,160, 1981 25. H.W. Brill-Edwards, U.S. Patent 3,801,353, 1974 26. R.L. Samuel and M.A. Lockington, Metal Treatment and Drop Forging, Vol 18, 1951, p 354, 407, 440, 495, and 543 and Vol 19, 1952, p 27 and 81 27. R. Drewett, Anti-Corros. Methods Mater., Vol 16, 1951, p 543 28. T.P. Hoar and E.A.G. Croom, Australas. Eng., Vol 63, 1950, p 56

29. S.C. Kung and R.A. Rapp, Oxid. Met., Vol 32, 1989, p 89 30. H.M.J. Mazille, Thin Solid Films, Vol 65, 1980, p 67 31. F.N. Mazandarany and R.D. Pehlke, Metall. Trans, Vol 4, 1973, p 2067 32. G.H. Meir, C. Cheng, R.A. Perkins, and W. Bakker, Surf. Coat. Technol., Vol 39/40, 1989, p 53 33. E. Godlewska and K. Godlewski, Oxid. Met., Vol 22, 1984, p 117 34. P. Galmiche, Metals Mater., Vol 2, 1968, p 241 35. P. Galmiche and A. Hivert, Canadian patent 806, 618, 1969 36. J.E. Restall and C. Hayman, U.S. patent 4,687,684, 1987 37. W. Johnson, K. Komarek, and E. Miller, Trans. Met. Soc. AIME, Vol 242, 1968, p 1685 38. B. Nciri and L. Vandenbulcke, J. Less Common Met., Vol 95, 1983, p 55 39. G.H. Marjinissen, Codeposition of Chromium and Aluminum during a Pack Process, High Temperature Protective Coatings, S.S. Singhal, Ed., TMS-AIME, Warrendale, PA, 1983, p 27 40. V.A. Ravi and R.A. Rapp, High Temperature Ordered Intermetallic Alloys III, C.T. Liu, A.I. Taub, N.S. Stoloff, and C.C. Koch, Ed., MRS Symposium Proceedings, Materials Research Society, Pittsburgh, PA, Vol 133, 1989, p 543 41. V.A. Ravi, P. Choquet, and R.A. Rapp, International Meeting on Advanced Materials, Materials Research Society, Pittsburgh, PA, Vol 4, 1989, p 483 42. V.A. Ravi and R.A. Rapp, Oxidation of High Temperature Intermetallics, T. Grobstein and J. Doychak, Ed., The Metallurgical Society, Warrendale, PA, 1989, p 543 43. W. Da Costa, B. Gleeson, and D.J. Young, J. Electrochem. Soc., Vol 141, 1994, p 1464 44. W. Da Costa, B. Gleeson, and D.J. Young, J. Electrochem. Soc., Vol 141, 1994, p 2690 45. W. Da Costa, B. Gleeson, and D.J. Young, J. Electrochem. Soc., Vol 88, 1996, p 165 46. R. Bianco and R.A. Rapp, J. Electrochem. Soc., Vol 140, 1993, p 1181 47. R. Bianco, R.A. Rapp, and J.L. Smialek, J. Electrochem. Soc., Vol 140, 1993, p 1191 48. J. Kipkemoi and D. Tsipas, J. Mater. Sci., Vol 31, 1996, p 6247 49. M.T. Kim, N.H. Heo, J.H. Shin, and C.Y. Kim, Surf. Coat. Technol., Vol 123, 2000, p 227 50. N.H. Heo, M.T. Kim, J.H. Shin, and C.Y. Kim, Surf. Coat. Technol., Vol 124, 2000, p 39 51. E. Fitzer, Arch. Eisenhuettenw, Vol 25, 1954, p 455 52. N.S. Gorbunov, Diffuse Coatings on Iron and Steel, The Academy of Sciences of the USSR, 1958, p 97 53. H.-W. Hsu and W.-T. Tsai, Mater. Chem. Phys., Vol 64, 2000, p 147

54. F. Fitzer and J. Schlichting, Coatings Containing Chromium, Aluminum, and Silicon for High Temperature Alloys, High Temperature Corrosion, R.A. Rapp, Ed., NACE International, 1981, p 604 55. M.A. Harper and R.A. Rapp, Oxidation of Metals, Vol 42, 1994, p 303 56. B.V. Cockeram, Surf. Coat. Technol., Vol 76–77, 1995, p 20 57. B.V. Cockeram and R.A. Rapp, Metal. Mater. Trans., A, Vol 26, 1995, p 777 58. B.V. Cockeram and R.A. Rapp, Mater. Sci. Engin. A, Vol 192/193, 1995, p 980 59. T.C. Munro and B. Gleeson, Metal. Mater. Trans., A, Vol 27, 1996, p 3761 60. M.T. Kim and J.S. Jung, Surf. Coat. Technol., Vol 161, 2002, p 218 61. Z.D. Xiang, S.R. Rose, and P.K. Datta, Mater. Chem. Phys., 9842, 2003, p 1


Selected References • • • • • • • • •

L.K. Bennett and G.T. Bayer, Pack Cementation Aluminizing of Steels, Surface Engineering, Vol 5, ASM Handbook, 1994, p 617–620 Corrosion Engineering, M.G. Fontana, 3rd ed., McGraw-Hill International Book Company, 1986 G.W. Goward and L.L. Seigle, Diffusion Coatings for Gas Turbine Engine Hot Section Parts, Surface Engineering, Vol 5, ASM Handbook, 1994, p 611–617 High Temperature Alloys for Gas Turbines, D. Coutsouradis, P. Felix, H. Fischmeister, L. Habraken, Y. Lindblom, and M.O. Speidel, Ed., Applied Science Publishers, London, 1978 High Temperature Corrosion, P. Kofstad, Elsevier Applied Science Publishers, Ltd., 1988 High Temperature Corrosion, R.A. Rapp, Ed., NACE International, 1981 High Temperature Protective Coatings, M. Khobaib and R.C. Krutenat, Ed., The Metallurgical Society, Warrendale, PA, 1987 Oxidation of High Temperature Intermetallics, T. Grobstein and J. Doychak, Ed., The Metallurgical Society, Warrendale, PA, 1989 Surface Engineering, R. Kossowsky and S.C. Singhal, Ed., Martinus Nijhoff Publishers, Pittsburgh, 1984

T. Mooney, Electroplated Coatings, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 772–785

Electroplated Coatings Revised by Ted Mooney, Finishing.com, Inc.

Introduction ELECTROPLATED COATINGS can be very effective in reducing the cost of corrosion, estimated in 2001 to be $276 billion, or 3.1% of the U.S. gross domestic product. Even when metallic coatings are applied to a substrate to provide decorative effects or to enhance material properties, such as wear or electrical contact resistance, the corrosion performance of the coating- substrate system must be carefully considered to avoid future degradation. Metallic coatings can be very useful in providing optimal properties for a part not obtainable if it were fabricated from a single material. The overall properties of coated parts are determined by the properties of the coating itself and by defects in the coating. The understanding of how defects in a coating affect the corrosion performance is critical in designing an effective coating system. Electroplating is the electrodeposition of an adherent metallic coating on an electrode to secure a surface with properties or dimensions different from those of the base metal (Ref 1). Electrodeposits are applied to metal substrates for one or more of the following reasons: decoration, protection, corrosion resistance, chemical inertness, wear resistance, buildup of substrate dimensions, electrical properties, magnetic properties, solderability, reflectance, and reduction of friction (Ref 2). For example, decorative and protective plating are combined in such applications as steel automobile bumpers, trim items made of zinc die castings or steel, trim items on large and small appliances, costume jewelry, electronic circuits, piano strings, construction and architectural hardware, plumbing fittings, electrical contacts, plastic items, and magnetic memory devices (Ref 2). This ability to be decorative as well as protective to substrates distinguishes electrodeposited coatings from metallic coatings applied by hot dipping or spraying. Metallic luster, achieved by bright plating or by polishing after plating, lends a distinctive appearance not provided by other types of coatings. However, corrosion resistance is the primary reason for the use of electrodeposited coatings. Electrodeposited coatings are also applied to protect substrate metals. Examples of this application include: • • • •

Tin, as well as chromium, plating of steel strip for food packaging and other container uses Electrogalvanizing, or zinc plating, of steel strip, sheet, stampings, forgings, wire, and screw machine products Zinc and cadmium plating of fasteners and other hardware items Chromium plating of gun bores

Electrodeposits are used to enhance properties other than corrosion protection. For example, many fasteners are coated with a bright cadmium or zinc deposit supplemented with a clear or colored chromate conversion film. This results in coatings that are both protective and decorative. Chromium plating, electroless nickel plating, and anodizing of aluminum (not actually an electroplating process, but an electrochemical process widely practiced at electroplating installations) impart very hard and exceptionally wear-resistant surfaces while simultaneously deterring corrosion. This article discusses the various factors that affect the corrosion performance of electroplated coatings, the effects of environment and the deposition process on substrate coatings, and the electrochemical techniques capable of predicting the corrosion performance of a plated part. Because the practical aspects of electroplating are covered in such publications as Surface Engineering, Volume 5 of the ASM Handbook (1994) and Ref 3, 4, 5, 6, this article is primarily concerned with the corrosion resistance of electrodeposited coatings. An exception is the discussion of mill- applied finishes, about which the open literature is sparse and limited in scope. Special attention is given to the design of coating systems for optimal protection of the substrate. An ideal electroplated

coating system is defined. Finally, attention is given to controlled weathering tests and accelerated tests used to predict and determine the relative durability of the coating.

References cited in this section 1. “Standard Definitions of Terms Relating to Electroplating,” B 374, Annual Book of ASTM Standards, American Society for Testing and Materials 2. C.L. Faust, “Electroplating,” Course 22, Metals Engineering Institute, American Society for Metals, 1983 3. F.A. Lowenheim, Ed., Modern Electroplating, 3rd ed., John Wiley & Sons, 1974 4. L.J. Durney, Ed., Electroplating Engineering Handbook, 4th ed., Van Nostrand Reinhold, 1984 5. F.A. Lowenheim, Electroplating, McGraw- Hill, 1978 6. M. Murphy, Ed., Metal Finishing Guidebook & Directory, Elsevier Science, 2001

T. Mooney, Electroplated Coatings, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 772–785 Electroplated Coatings Revised by Ted Mooney, Finishing.com, Inc.

Idealized Coating Coatings are either anodic or cathodic to the substrate. Only a very rough idea of which situation applies can be obtained from the electromotive force (emf) series (Table 1). Coatings that are anodic to a substrate in one environment may be cathodic to the same substrate in a second environment. The environment itself may change from the surface of the coating to one that exists within any defects, such as scratches, cracks, or corrosion pits. These defects are particularly important if the coating is cathodic to the substrate; in this case, if the substrate becomes exposed, rapid corrosion of the substrate may occur as it attempts to protect the coating. If the coating is anodic to the substrate, defects are somewhat less important, although they still must be considered. Thus, if the substrate becomes exposed, it will be protected by the corrosion of the coating. However, if the corrosion should proceed to such an extent that a large area of the substrate is exposed, the coating will lose its effectiveness, because the electrical resistance of the substrate limits the area that the coating can protect. Table 1 Electromotive force series of metals and alloys Solution concentration: 1 N; temperature: 25 °C (75 °F) Metal Ion Standard reduction potential (E0), V 2+ Ca -2.87 Calcium Na+ -2.71 Sodium Magnesium Mg2+ -2.37 Aluminum Al3+ -1.66 Zn2+ -0.763 Zinc Chromium Cr3+ -0.71 Cr6+ -0.56

Iron Cadmium Nickel Tin Lead Hydrogen Copper

Fe2+ Cd2+ Ni2+ Sn2+ Pb2+ H+ Cu2+ Cu+ Ag+ Pt4+ Au+

-0.44 -0.40 -0.25 -0.136 -0.126 0 0.337 0.521 0.799 0.90 1.68

Silver Platinum Gold Source: Ref 2 The nature of any corrosion products formed also must be taken into account. For example, in industrial environments, zinc is more protective of steel than cadmium is because the zinc sulfate (ZnSO4) corrosion product formed has a lower solubility than cadmium sulfate (CdSO4); ZnSO4 thus provides a blocking action to reduce subsequent corrosion. This situation is reversed in marine environments, in which the corrosion products are more likely to be carbonates and chlorides. Zinc produces more soluble corrosion products than cadmium does in this environment. The most desirable situation is to have the coating anodic to the material beneath it and at the same time to have a very low corrosion rate. This general rule has been exploited to a high degree of sophistication by the automotive industry. Chromiumplated exterior steel parts have a coating system consisting of several layers of metals, each of which is anodic to the layer beneath it. There are accurate lifetime predictions based on the electrochemical potential differences and the thicknesses of the layers. Multilayer corrosion standards that ensure that the coatings meet specifications are available from the National Institute of Standards and Technology (NIST). Even the chromium layer has been designed to be microcracked or microporous to spread the corrosion reactions over a large area. Corrosion is then distributed over a large area, and localized pitting is minimized. This is discussed in more detail in the section “Applications for Electrodeposited Coatings” in this article.

Reference cited in this section 2. C.L. Faust, “Electroplating,” Course 22, Metals Engineering Institute, American Society for Metals, 1983


Substrates The emf series is a table that lists in order the standard electrode potentials of specified electrochemical reactions (Ref 1). The potentials are measured against a standard reference electrode when the metal is immersed in a solution of its own ions at unit activity (Table 1). Similarly, the galvanic series is a list of metals and alloys arranged according to their relative potentials in a given environment (Ref 1). Figure 1 shows a galvanic series of commercially used metals and alloys in seawater. The positions of metals and alloys in a galvanic series—and the corrosion behavior of metals and alloys—depend on the specific environment to which the materials are exposed. Therefore, there are as many galvanic series as there are aqueous environments. Generally, the relative positions of metals and alloys in both emf and galvanic series are the same. An exception is the position of cadmium with respect to iron and its

alloys. In the emf series, cadmium is cathodic to iron, but in the galvanic series (at least in seawater) cadmium is anodic to iron. Thus, if only the emf series were used to predict the behavior of a ferrous metal system, cadmium would not be chosen as a sacrificial protective coating, yet this is the principal use for cadmium plating on steel.

Fig. 1 Galvanic series of metals and alloys in seawater. Alloys are listed in order of the potential they exhibit in flowing seawater; those indicated by the black rectangle were tested in low-velocity or poorly aerated water and at shielded areas may become active and exhibit a potential near -0.5 V. Source: Ref 7 The importance of knowing the position of substrates in the emf and a specific galvanic series is frequently overlooked. The sole consideration is often the position of the protective coating in these series. It is prudent, however, to select a coating as close as possible in potential to the substrate in moist environments, because few coatings are completely free of pores, cracks, and other defects. Even if a coating does start out defect free, scratching, normal wear, or gouging can change this condition. Accordingly, in a corrosive environment, plated components present at least two (not one) metal to the hostile galvanic-corrosion forces (see the section “Electrochemical Predictions of Corrosion Performance” in this article).

Cleanliness of the Substrate Regardless of the composition and metallurgical condition of a substrate, its surface must be free from soils and oxides before being plated. The ASTM International Committee on Selection of Cleaning Processes classifies soil into six groups: pigmented drawing compounds, unpigmented oil and grease, chips and cutting fluids, polishing and buffing compounds, rust and scale, and miscellaneous surface contaminants, such as lapping compounds and residue from magnetic- particle inspection. Cleaning methods are discussed in Ref 2, 3, 4, 5, 6. Although these methods are beyond the scope of this article, the point is made that unclean substrates often defeat the purpose of electrodeposited or other plated coatings. Freedom from water break after the activating rinse and before immersion of a part in the plating tank is an essential prerequisite for satisfactory plating. Water break is defined as the appearance of a discontinuous film of water on a surface, signifying nonuniform wetting and usually associated with surface contaminations (Ref 1). A false reading of freedom from water break often occurs if the inspection is made after the rinse following a surfactant-laden alkaline cleaner. This misleading conclusion is nullified by inspection after acid dipping and rinsing.

Knowledge of the Nature of Substrates Before designing or applying a protective coating to a metal, the plater must know the nature and essential properties of the substrate. Generally, a substrate is known to be ferrous or nonferrous, but these generic terms are inadequate. For example, there are thousands of ferrous metals and alloys, ranging from cast iron (high carbon) to ingot iron (very low carbon) and the entire range of carbon and alloy steels. Also, the microstructure, hardness, yield and tensile strengths, and internal stress of the substrate have major influences on the protective quality and durability of plated coatings. High-strength, high-carbon, high-hardness steels are susceptible to hydrogen embrittlement when subjected to alkalies, acids, and plating processes. Aluminum and zinc alloys and, to a lesser extent, chromium are amphoteric (soluble in both acidic and alkaline solutions). Accordingly, these metals must be treated with inhibited solutions to prevent chemical attack. Methods of determining the nature and condition of the substrates are described in the following sections. Information on the Substrate. A design engineer, draftsman, engineer, or project manager in an organization with in-house plating facilities should have detailed information on the analysis, microstructure, and mechanical properties of the substrate. In a plating job shop, the customer should have someone informed on the subject. The plater is advised to ask these potential sources for as much information about the substrate as can be obtained. Engineering Drawings and Associated Lists. Study of engineering drawings, parts lists, bills of materials, and pertinent manuals will usually clear up questions concerning the substrate submitted for plating. The bill of material, for example, may simply specify 1095 steel. This informs the plater that he is dealing with a highcarbon (0.95% C) and steel that can be heat treated to a high degree of hardness (for example, 50 to 60 HRC). This alerts the plater to be concerned with hydrogen embrittlement. As another example, if a drawing or associated list specifies aluminum alloy 6061-Tx, the plater can deduce that the substrate is heat treatable by precipitation hardening to various tempers (degrees of hardness). Such an alloy is subject to developing a smut or to leaving a heavy oxide layer after being alkaline etched or cleaned. An experienced plater will recognize this and will know to treat the cleaned and etched part in a desmutter/deoxidizer in the treating sequence before undertaking the plating process.

Chemical Spot Tests. A simple chemical spot reaction test is documented in Ref 6 and 7. This method is based on the use of a magnet and five chemicals. Three of the chemicals are different concentrations of nitric acid (HNO3), another is concentrated hydrochloric acid (HCl), and the fifth is a 10% solution of potassium ferricyanide (K3Fe(CN)6). Although it tests for only one element at a time, this simple test distinguishes among a large number of commonly used steels and nickel alloys. Electrolytic Spot Test Method. This is a commercially available method, complete with chemicals, a power source, and instructions. The operation is divided into four steps. First, a test paper moistened with a conductive solution is placed in contact with the unknown metal. Second, using a direct-current (dc) power source with the unknown metal connected to the anode terminal (+) and a metal probe to the cathode terminal (-), a small amount of the unknown metal is dissolved into the test paper. Dissolution is brought about by electrolytic action when the dc current is switched on. Third, a few drops of a specific color developer are added to the wet spot on the paper. The developer reacts chemically with a characteristic component of the dissolved metal. Fourth, almost immediately, a color appears on the paper if a specific suspected alloying element is present in the unknown. This method is semiquantitative because the relative color intensity is proportional to the amount of alloying element or major metallic constituent. By comparing the results of the unknown with those of standard alloy samples, the nature of the unknown substrate can be adequately identified. Electronic Alloy Identification. This method is quite rapid and depends on the Peltier Effect (also the basis of the thermocouple). If a current of electricity is sent through a circuit containing two dissimilar metals, heat is emitted at one junction and absorbed at the other junction. A characteristic voltage (potential) is generated at the thermocouple junction that causes an electrical current to flow. This voltage varies with the metals that constitute the couple. The electronic alloy identification meter operates on this principle. It measures the voltage generated by the action of a file of known composition scratched along the surface of an unknown alloy. The file acts as the positive meter input probe, with a clip-on providing negative reference. In a commercially available apparatus, the thermocouple voltage is shown on a battery-powered digital display. This voltage is characteristic of specific alloys or classes of alloys. Again, the instrument can be calibrated against known alloy specimens that are sold as standards.

Physical and Mechanical Properties of Substrates The susceptibility to corrosion of a plated substrate depends on physical properties and mechanical condition. For example, surface profile (roughness, presence of scratches, pits, lapped metal, and so on) can profoundly affect the durability of plated ware. Some of the harmful effect can be avoided by mechanical or electropolishing, polishing and buffing, shot peening, or glass beading. Internal stresses can also be harmful. These may dictate heating for stress relief or shot peening or glass beading to convert surface tensile stresses to compression. The physical properties of a substrate, particularly at the surface, are often as important to the metal finisher as its composition. Porosity, hardness, melting point, coefficient of thermal expansion, and permeability to hydrogen (nascent, or atomic, and gas) are vital to many operations. Porosity of substrates, particularly in castings and sintered powder metal compacts, causes many problems because of the difficulty of rinsing processing solutions. Capillary action tends to hold solutions with forces that defy normal rinsing efforts. This results in spotting out of corrosive liquids, which appear as stains and corrosion products on the plated surface. It also causes contamination of processing baths. Some surface irregularities of substrate can be compensated for by the leveling properties of specially formulated plating solutions. Leveling action is the ability of a plating solution to produce a surface smoother than that of the substrate. The property of the plating solutions that produces leveling is also referred to as microthrowing power. Another property of substrates that can cause difficulties is susceptibility to hydrogen absorption. High-carbon and, particularly, hardened steels and titanium are susceptible to embrittlement when processed by methods that subject them to atomic or gaseous hydrogen. Recognizing this weakness of such substrates, the designer or specifier of plating sequences must select alkaline cleaning and oxide removal (pickling agents, activators) procedures that lessen or avoid hydrogen evolution. Mechanical cleaning and mechanical plating (also known as peen plating) are examples of the methods used to avoid hydrogen embrittlement. Because total avoidance of hydrogen evolution is often not possible, a step or steps must be added to the plating sequence to remove hydrogen. This usually consists of baking plated parts at low temperatures (205 °C, or 400 °F) to release absorbed hydrogen. Time and temperature is also an important variable in ensuring the release of hydrogen and restoration of the mechanical properties of the plated substrate.

Substrate metals that have been heavily worked by such operations as deep drawing, swaging, polishing and buffing, grinding, machining, forging, and die drawing often come to the plating shop with a damaged layer on the surface. This heavily worked layer is termed a Beilby Layer in honor of Sir George T. Beilby, who stated that this layer was amorphous or vitreous rather than crystalline. He compared this weak, somewhat brittle layer to substances that resemble the glasslike form assumed by the silicates when they are solidified from the molten state (Ref 8). X-ray analyses in later years proved that severe cold working causes fragmentation of crystals into particles of different orientations, but that the surface metal is not vitreous or amorphous. However, this smeared, shattered layer is inherently low in ductility and fatigue strength and is therefore a weak foundation for plated coatings. This is particularly true of metal objects that require a surface of substantial corrosion resistance and high mechanical properties as well as metal objects that will be further formed or distorted in service. To avoid the weakness introduced by such layers, abrasive cleaning, machining, or electrolytic polishing are required to remove them before plating.

Stresses in Substrates Substrates with locked-in stress reduce the corrosion resistance of electroplated coatings relative to the same coating on an annealed substrate. Stressed metal is anodic to annealed or lesser-stressed metals and is therefore more prone to corrosion in unfavorable service conditions. Substrates with flaws on the surface, stress-raising or stress-localizing scratches, or abrupt changes in section are less resistant to corrosion than smooth, streamlined parts. The coefficients of thermal expansion of the coating and substrate should also be matched as closely as possible to minimize stresses caused by differential expansion during temperature cycling.

Stress-Corrosion Cracking of Substrates Stress-corrosion cracking (SCC) is a cracking process that requires the simultaneous action of a corrodent and sustained tensile stress. Certain alloys are susceptible to SCC when subjected to certain corrosive environments. Examples are copper alloys, such as Cu-30Zn (cartridge brass), and nickel-chromium alloys. Ammoniacal materials, such as aqueous ammonia (NH4) and amines, produce such failures in stressed copper alloys. Halides, such as chlorides and bromides, attack certain nickel-chromium alloys with this type of corrosion. The basic characteristics of SCC are summarized in Ref 9. There is no available evidence that plated coatings on substrates susceptible to SCC will prevent its occurrence. Stress relief of susceptible alloys is often adequate for avoiding SCC. For example, brass cartridge cases, which are made by a deep-drawing process (leaving high residual stress), are stress relieved in a temperature range that produces recovery without recrystallization. Stress relieving at a higher temperature would produce recrystallization and an unacceptable reduction in mechanical properties. Avoidance of the use of SCC-susceptible alloys in known hostile environments is another approach to avoiding this mode of failure, but this is not always feasible. For example, for the brass cartridge case cited previously amine stabilizers in the smokeless powder propellant obviated control of the environment; therefore, stress relief was required. However, use of SCC-susceptible metals in corrosioncracking environments should be avoided whenever possible.

References cited in this section 1. “Standard Definitions of Terms Relating to Electroplating,” B 374, Annual Book of ASTM Standards, American Society for Testing and Materials 2. C.L. Faust, “Electroplating,” Course 22, Metals Engineering Institute, American Society for Metals, 1983 3. F.A. Lowenheim, Ed., Modern Electroplating, 3rd ed., John Wiley & Sons, 1974 4. L.J. Durney, Ed., Electroplating Engineering Handbook, 4th ed., Van Nostrand Reinhold, 1984 5. F.A. Lowenheim, Electroplating, McGraw- Hill, 1978 6. M. Murphy, Ed., Metal Finishing Guidebook & Directory, Elsevier Science, 2001

7. R.S. Treseder, Ed., The NACE Corrosion Engineer's Reference Book, National Association of Corrosion Engineers, 1980 8. G.T. Beilby, The Hard and Soft States in Metals, J. Inst. of Met., No. 2, 1912, p 149–185 9. R.G. Baker, “Corrosion of Electrodeposited Coatings,” Slide Lecture 46, American Electroplaters and Surface Finishers Society, 1981


Design for Plating Much can be done at the design stage to improve the plateability of metal stampings, castings, forgings, and extruded shapes. In doing so, the corrosion protection afforded by the plating can be enhanced. Corrosion resistance of a plated object depends on coating thickness and distribution, and these factors depend largely on the design of the parts. ASTM B 507 (Ref 10) covers practices for the design of articles to be electroplated on racks with nickel, but its significance goes beyond nickel plating. It also applies in most respects to all rackplated coatings. The general principles of design are discussed in Ref 10, which states that the electrodeposition of one metal onto another follows a pattern such as that exemplified in Fig. 2, showing current distribution and the distribution of the resultant electrodeposit (Fig. 2).

Fig. 2 Current distribution (a) and effect of substrate geometry (b) on thickness of electroplated nickel deposits. For the 120° substrate in (b), the ratio of average coating thickness to minimum coating thickness is 1.9; for the 90° substrate, 2.7; and for the 60° substrate, 3.3. Source: Ref 10 Distribution of the deposit will vary slightly as a function of bath type, additives, current density, and temperature. The coating is preferentially deposited at protuberances, sharp corners, edges, fins, and ribs on the component being plated; to avoid coating buildup, these areas should be rounded with a minimum radius of 0.4 mm (0.015 in.). A 0.8 mm (0.03 in.) radius is preferable. Conversely, the base of any cavity, recess, or depression in the component being plated will be deficient in plate thickness; these features should be designed

to avoid sharp fillets. In general, such features should be designed so that the ratio of width to depth of the depression is held to a minimum of 3, and the minimum radius at the edge and base of an indentation should be one-fourth of the depth of the indentation. When sharp angles are required for functional purposes, the electroplater cannot be expected to meet minimum coating thickness requirements in these areas and corrosion resistance may be inadequate. When designs cannot be changed, electroplaters resort to one or more of the following methods to overcome the aforementioned difficulties: provision of inside anodes, use of thieves or robbers (nonconductive shields that slow or prevent deposition) to lower the cathodic current density at protuberances, bipolar electrodes, changing electroplating baths to ones with greater covering power, and switching from electroplating to autocatalytic (electroless) plating. This last method is widely used on parts that cannot be properly plated by electrodeposition. The limitation is that the most prevalent coating applied autocatalytically—electroless nickel—deposits an alloy of nickel and phosphorus (phosphorus content varies with the solution used and the deposition condition), not unalloyed nickel. There is also an economic disadvantage in switching to these processes. However, in most cases in which electroless processes have been adopted, they have been specified for reasons other than merely to compensate for part designs that cannot be readily electroplated with uniform coatings. An article should not be designed for plating, but rather to perform its intended function (Ref 11). If the part is successfully designed, provision will have been made for corrosion resistance, hardness, lubricity, decorative appeal, or whichever required property the finish process is to supply. Function, fit, form, and aesthetics—the major dictates of design criteria—often take precedence over the practicality of corrosion prevention by plated coatings; this need not be the case. Design manuals must be instructional with regard to the effects of geometry on function, and designers should be taught these principles. Additional information on design for electroplating is available in Ref 12. Table 2 lists proper and improper designs for various features of electroplated components. Plating of relatively small parts in rotating containers is termed barrel plating. This differentiates it from rack plating, in which individual parts are held in a rack or fixture during plating. Fastening hardware, such as nuts, bolts, washers, screws, and studs, are barrel plated, as are numerous critical electrical and electronic components. Many of these parts are as important from the corrosion resistance standpoint as rack-plated parts are—perhaps more so because such parts as fasteners and connectors are used in, and therefore coupled to, other materials. This raises the problem of galvanic corrosion. Table 2 Influence of substrate design features on electroplateability

Bulk parts in a plating barrel are subject to a greater variation in cathodic current density than parts plated on racks. Cathode contact is made by danglers, chains, buttons, or studs near the bottom of the load, while the plating occurs only at the surface of the load. This produces a large statistical variation in thickness of the electrodeposits on barrel-plated parts; this is not the case for rack-plated parts. Corrosion of a plated part is often controlled by the minimum thickness of protective coating; therefore, if parts are to be barrel plated, special design practices should be observed, including (Ref 12): • • • • • •

Avoid blind holes, recesses, and joint crevices that can retain tumbling compounds and metal debris from previous operations. Avoid intricate surface patterns that will be blurred by barrel finishing. Design parts to withstand the multiple impacts of barrel rotation. Design small, flat parts with ridges or dimples to prevent nesting. Design for good entry and drainage of solutions during rotation by using simple shapes. Design surfaces to be plated so that they undergo proper mechanical preparation and cleaning and to ensure that they receive their share of metal deposit. They should be convex, if possible, rather than recessed.

In a larger sense, the design of parts for plating parallels good design practice for corrosion resistance. Avoidance of sharp edges, burrs, or other areas of high free energy, avoidance of abrupt changes in section, recesses, and other traps for moisture, and avoidance of scratches, notches and other anodic-site producers are common design criteria for parts in general as well as for plated parts. More information on designing to minimize corrosion is available in the article “Designing to Minimize Corrosion” in this Volume.

References cited in this section 10. “Standard Practice for Design of Articles to be Electroplated on Racks with Nickel,” B 507, Annual Book of ASTM Standards, American Society for Testing and Materials 11. J. Hyner, Design for Plating, Electroplating Engineering Handbook, L.J. Durney, Ed., Van Nostrand Reinhold, 1984, p 50 12. Quality Metal Finishing Guide, Metal Finishing Suppliers' Association


Advantages and Disadvantages of Electroplated Coatings Alternatives to electroplating for the protection of metal surfaces against corrosion include organic finishing, hot dip coating, metal spraying, chemical vapor deposition (CVD) and physical vapor deposition (PVD), vacuum deposition by sputtering and ion implantation, and ceramic coating. (See the articles “Organic Coatings and Linings,” “Continuous Hot Dip Coatings,” “Batch Process Hot Dip Galvanizing” “Thermal Spray Coatings,” “CVD and PVD Coatings,” “Surface Modification Using Energy Beams,” and “Chemical-Setting Ceramic Linings” in this Volume). The decision to use one or more of these methods must consider at least the following factors: • • • •

Applicability to the substrate in question (For example, zinc plating is applicable as a protective coating to steel, but not cast iron.) Purposes of the finish (Is color a requirement? Will the object to be finished fit into available plating equipment?) Cost of suitable electroplating versus alternative methods of coating the part (What finish will be functionally adequate and applicable at the least cost?) Anticipated service conditions and length of service desired under these conditions (SC5, extended very severe; SC4, very severe; SC3, severe; SC2, moderate; SC1, mild) (Ref 13)

Advantages of Electroplating Variety of Coatings. The principal metals that are widely plated are cadmium, zinc and its alloys, chromium, copper, gold, nickel, tin, and silver. Other metals that are plated by in-house shops and some job platers are cobalt, indium, iron, lead, palladium, platinum, rhenium, and rhodium. Plateable alloys add versatility to the catalog of plated coatings. The commonly plated alloys are brass, bronze, copper-cadmium, tinlead (solder), tin-zinc, gold alloyed with many metals, and nickel alloys with iron and cobalt. Zinc alloy platings, including zinc-cobalt, zinc- iron, zinc-nickel, and tin-zinc are of ever-growing importance. Ternary alloys, such as karat gold alloys and Cu-Sn-Zn, are also available. Nickel-phosphorus and nickel-boron are readily plated by the autocatalytic (electroless) process. Nickel-phosphorus can also be electrodeposited. Coatings of many types can be codeposited with occluded particles by electrodeposition or electroless deposition to modify friction and wear properties. Particles used for these applications range from deposited or occluded abrasives such as diamond, silicon carbide (SiC) to lubricious materials such as polytetrafluoroethylene (PTFE). Versatility of Application. Plated coatings can be applied to racked parts, in barrels (both by immersion in tanks and manually), semiautomatically, or on microprocessor-controlled fully automatic machines. Local, selective

application of a coating, if necessary, can be accomplished by such methods as brush plating (also known as electrochemical metallizing and tampon plating). Similarly, precious metals, such as gold and palladium, have been conserved in electronics applications by selective plating on continuous strips. This is referred to as reelto-reel plating. The adaptability of plating methods to the job at hand is limited only by the imagination and skill of the plater. Availability of Applicators. Plating job shops and in-house applicators are plentiful in the United States and throughout the world. The technology is popularized and nurtured by such technical societies and trade associations as the American Electroplaters and Surface Finishers Society, the National Association of Metal Finishers, the Metal Finishing Suppliers' Association, and the Society of Automotive Engineers Airline Plating Forum. The electrical conductivity of electrodeposited and other metallic coatings gives them an advantage over most organic and ceramic-base coatings in applications that require this characteristic. Applications in which electrical conductivity is necessary include the welding, electrical bonding, or grounding of objects as well as electrical and electronic circuitry. Galvanic Protection of Substrates. Zinc and cadmium deposits are anodic to steel and protect steel substrates by sacrificial action in many wet environments. In aircraft and other aerospace applications, cadmium plating is used on steel components in contact with aluminum alloys because cadmium is closer in potential to aluminum than steel is (Fig. 1). On coil-coating lines for prepainted steel sheet and strip, electrogalvanizing stages are provided in the wet section to protect the substrate at points of failure of the organic overcoating. Electrogalvanized steel, often further treated with chromate conversion coatings, is used for various purposes because of the sacrificial protective value of the zinc coating. Duplex or triplex coatings of nickel, often with a final thin coating of chromium, are another example of a more complex galvanic protective system. Weldability and Solderability. Metal deposits are amenable to welding and soldering—an advantage over most organic coatings and porcelain enamel. Certain coatings are more weldable and solderable than others, which leads to the application of special types of plating to facilitate a specific type of joining or fabricating method. Control of Thickness. Unlike hot dip coating, electrodeposition affords a means of predicting and controlling the amount of metal deposited. This is possible through the validity and knowledge of Faraday's laws, which are as follows (Ref 2): • •

The quantity of any element discharged at an electrode is proportional to the quantity of electricity that is passed. The quantities of different ions of elements discharged by a given amount of electricity are proportional to the electrochemical equivalents of those elements.

The quantity of electricity is measured in Coulombs (C) (C = A · s), and the electrochemical equivalent of an element is its atomic weight divided by its valence. It is known that 96,480 C, or 1 Faraday will deposit 1 g equivalent weight of any element at 100% electrode efficiency. Electrode efficiencies are a function of plating electrolytes and conditions. They can be determined by coulometric measurement and are published for most plating processes. Because most plants control plating by measuring ampere hours, the amount in grams of metal plated, Md, is calculated by the relationship:

where EEM is the electrochemical equivalent of metal M (atomic weight of M divided by valence of M), and CEM is the cathode efficiency of the plating bath and conditions. Ductility and Formability. Substrate metals coated by electrodeposition can be formed more readily than some metals coated by hot dipping or certain organic coatings. They do not result in intermetallic alloy layers of the type common to ordinary hot dip galvanizing, for example. Diffusion of some electrodeposits into substrates, and vice versa, occurs after plating as a function of temperature, time, and the diffusion tendencies of the metals involved. These diffused layers are normally thin and require protracted periods of time to develop, usually after the component has been formed to a desired shape. Electroless Plating. The coating of the interiors of tubes or other recessed areas uniformly and with controlled thickness is the special province of autocatalytic, or electroless, plating. Mechanical Plating. Mechanical, or peen, plating is a method of coating substrates, principally steel, with malleable metals, such as zinc, cadmium, tin, or their alloys. In mechanical plating, hard, small, spherical objects, such as glass shot, are tumbled against the parts to be plated in the presence of finely divided metal

powder, such as zinc dust, and appropriate chemicals. Special advantages of this method over electroplating are the relative absence of problems with hydrogen embrittlement, the simplicity of operation with small parts in a single rotating container, and the ability to deposit desirable alloys that cannot be electroplated. As an example of the latter, tin and cadmium alloys have been found to provide excellent protection of steel against corrosion in marine environments.

Disadvantages of Electroplating Color Limitations. The colors of electrodeposited coatings are limited to copper, silver metallic, brass, bronze, and gold and its alloys; black, iridescent, or olive drab chromate conversion coatings; and patinas formed on copper and its alloys by chemical and electrochemical treatments. Colored Anodic Coatings on Aluminum. Anodic oxidation of aluminum, an electrochemical frequently employed at electroplating shops, provides a protective coating that can be colored. The coating is integral to the substrate and can be colored by dyeing, integral color anodizing, or electrolytic coloring (Ref 6, 14). These coatings are highly resistant to abrasion and corrosion and are widely used on architectural forms. More information on these types of coatings is available in the article “Aluminum Anodizing” in this Volume. Inhospitable Substrates. Some substrates, such as cast iron, and certain substrate conditions, such as porosity, require special plating procedures. Unimpregnated powder-metal compacts and porous castings that can be plated are subject to spotting out, or bleeding, because of retention in pores of cleaning and plating alkalies, acids, and salts. Limitations of Design. Parts are designed to function in the most economical manner. This is not always compatible with plating. Where the geometry of a part is such that cathodic current distribution will be unfavorable to uniform coatings, it is sometimes wise to seek an alternative protective coating. The criteria here are that the coating must achieve the required minimum thickness on significant surfaces, must meet all functional and dimensional requirements, and must accomplish these objectives at a reasonable cost. Size Limitations. Large metal structures are often beyond the capabilities of electrodeposition. The limitation lies in the capacity of tanks to contain the electrolyte. Exceptions are certain architectural structures and statuary, which may be brush plated. Such coatings are almost always on copper or bronze and consist of nickel topped with gold electrodeposits. Steel structures, such as bridges, dams, and building elements, are more advantageously hot dip galvanized and painted, or painted without zinc but with specially designed organic coating systems. Environmental Issues. Electroplating as a whole has been under enormous environmental pressures worldwide since the late 1970s. Intrinsic to virtually all electroplating processes is a requirement for washing the plating solution off of the ware with copious volumes of fresh rinse water, which then presents the operator with the problem of dealing with large volumes of water contaminated with salts of heavy metals. It is expensive and often impractical or impossible to return these diluted plating solutions to the process tanks, while also unacceptable to discharge them untreated to surface waters or sanitary sewer systems. Similarly, it is inherent in the process that cleaning solutions, activating solutions, posttreatment dips, and even some plating solutions become exhausted or hopelessly contaminated and require treatment and disposal to a secure landfill. Many plating operations require mist evacuation, which holds the potential of polluting the air unless adequate fume scrubbing is ensured. None of these problems are insurmountable, but the triple threat of the potential for contaminating water, land, and air has made pollution avoidance an expensive and knotty problem for plating practitioners. Additionally, some of the individual plated metals and processes present problems. Cadmium is now recognized as a biocumulative poison with toxic effects similar to those of mercury and lead; cadmium plating is expressly forbidden on most or all parts in many countries. Nickel plating is no longer permitted in jewelry applications in much of the world. The hexavalent chromate dips traditionally used on zinc- plated surfaces to forestall white rust are now forbidden in enough of the world that they are no longer used anywhere in the manufacture of automobiles.

References cited in this section 2. C.L. Faust, “Electroplating,” Course 22, Metals Engineering Institute, American Society for Metals, 1983

6. M. Murphy, Ed., Metal Finishing Guidebook & Directory, Elsevier Science, 2001 13. “Standard Specification for Electrodeposited Coatings of Copper Plus Nickel Plus Chromium and Nickel Plus Chromium,” B 456, Annual Book of ASTM Standards, American Society for Testing and Materials 14. S. Wernick and R. Pinner, The Surface Treatment and Finishing of Aluminum and Its Alloys, Robert Draper, 1972


Effects of Deposition Parameters The process used to apply the coating may have an important effect on the properties of the entire coating system. One of the more important considerations is the effect of hydrogen formation during electroplating. Hydrogen may embrittle the substrate. This is particularly important when cadmium, chromium, or zinc is used as a coating over high-strength steel. Atomic hydrogen can combine to form high-pressure hydrogen gas at defects or regions of poor adhesion, causing blisters in the coating. Appropriate standards have been developed to cover relief from hydrogen embrittlement in high-strength steels (Ref 15). Hydrogen can also be introduced into the coating system during one of the cleaning or pickling steps conducted before electroplating.

Reference cited in this section 15. “Method for Mechanical Hydrogen Embrittlement Testing of Plating Processes and Aircraft Maintenance Chemicals,” F 519, Annual Book of ASTM Standards, American Society for Testing and Materials


Mechanisms of Corrosion Prevention of Plated Coatings Coatings applied by electrodeposition or electroless plating protect substrate metals in one of three ways. The first is cathodic protection of the substrate by sacrificial corrosion of the coating—for example, zinc and cadmium coatings are electronegative (or anodic) to steel. Second is barrier action, that is, interposing a more corrosion-resistant coating between the environment and a less corrosion-resistant substrate. Examples of this are Cu-Ni-Cr and Ni-Cr coating systems over steel and zinc alloy automotive parts. Third is environmental

modification or control in combination with a nonimpervious barrier coating. An example of this type of protection is the electrolytic tinplate used in food packaging. Anodic Coatings. Zinc and cadmium deposits will protect steel at pores and other discontinuities by cathodic protection. Observation of the emf series (Table 1) shows that zinc is electronegative to steel in potential when these metals are immersed in solutions of their own ions. Under the same conditions, cadmium is more noble than iron, but in actual service under several environmental influences, cadmium is anodic to iron and is therefore sacrificially protective. Figure 3 shows the galvanic protection offered by a zinc coating to a steel substrate. This mechanism of galvanic protection gives rise to the term galvanized steel, which refers only to zinc-coated steel. This old, but commonly used term is modified to electrogalvanizing when referring to electrodeposited zinc in order to differentiate it from hot dip galvanizing.

Fig. 3 Principles and mechanism of galvanic protection of a substrate by a coating. (a) Galvanic protection of a steel substrate at a void in a zinc coating. Corrosion of the substrate is light and occurs at some distance from the zinc. (b) Galvanic protection of a steel member (1) by a zinc member and by an impressed electric current flowing from an external dc power source (E) using an insoluble platinum electrode (2). Source: Ref 2 Cathodic Coatings. With reference to Table 1, it is apparent that the metals commonly plated, other than zinc and cadmium, are electropositive (more noble or cathodic) to steel. This is also evidenced by Fig. 1. Such coatings on steel are therefore expected to act strictly as barriers to prevent corrosion of the steel substrate. For this to be successful, the coatings must be pore free and flaw free. It is very difficult to obtain impervious coatings of more noble metals on steel or zinc alloys. Substantial metal thicknesses, usually 25 μm (1 mil) or more, are required. With some of these metallic deposits, particularly the noble metals, thinner coatings applied by pulse plating are more impervious than conventional plated metals, but this is the exception. Generally, electroplated or hot dipped coatings that are completely free of pores and other discontinuities are not commercially feasible (Ref 2). Pits eventually form at coating flaws, and the coating is penetrated. The substrate exposed at the bottom of the resulting pit corrodes rapidly (Fig. 4). A crater forms in the substrate, and because of the large area ratio between the more noble coating and the anodic crater, the crater becomes anodic

at high current density. Electrons flow from the substrate to the coating as the steel dissolves. Hydrogen ions (H+) in the moisture accept the electrons and, with dissolved oxygen, form water at the noble metal surface near the void. Use of an intermediate coating that is less noble than a surface coating but more noble than the base metal may result in the mode of corrosion shown in Fig. 5. This would be typical of a costume jewelry item with the substrate being brass, the intermediate coating nickel, and the topcoat tarnish-resistant gold. It is also exemplified by nickel-chromium coating systems.

Fig. 4 Crater formation in a steel substrate beneath a void in a noble metal coating, for example, passive chromium or copper. Corrosion proceeds under the noble metal, the edges of which collapse into the corrosion pit. Source: Ref 2

Fig. 5 Corrosion pit formation in a substrate beneath a void in a duplex noble metal coating. The top coating layer (M1) is cathodic to the coating underlayer (M2), which is in turn cathodic to the substrate (M3). As in Fig. 4, the coating tends to collapse into the pit. Source: Ref 2 Exceptions to Predictions. The durability of a plated coating system cannot always be accurately predicted by conventional analysis of the potentials shown in the emf or galvanic series. Much depends on environmental exposure and the oxidized films formed in pores and discontinuities of plated coatings. For example, some corrosion specialists predicted copper-plated zinc one-cent pieces would be unsuitable coinage because of rapid corrosion. The coating is barrel-plated copper on a zinc-rich alloy, with the coating thickness varying from 4 to 8 μm (0.15 to 0.3 mil). In this thickness range, the coating is far from impervious, or pore free. In the 1970s, before the minting of these coins, pictures were published showing exfoliated, voluminous zinc corrosion products on sample pennies after exposure to the ASTM B 117 neutral salt-spray test (Ref 16). The predicted result did not occur after more than 10 years of service. This indicates that the service conditions to which these coins were subjected are more forgiving than a salt-spray test. Another example is the case of lead-plated steel, which is quite durable despite the fact that lead is more noble than iron. This is the case because a lead corrosion product develops in the pores (most likely lead sulfate PbSO4) and arrests the progress of corrosion. Importance of Thickness of Plated Coatings. In general, the thicker a metallic coating is, the less porous it is and the better it serves as a barrier against corrosion. Therefore, the measurement of thickness is a most important feature of testing electrodeposited and electroless coatings. Table 3 indicates the applicability of coating thickness measuring methods. American Society for Testing and Materials (ASTM) and International Standards Organization (ISO) standards covering thickness measurement methods are given in Ref 17.

Table 3 Applicability of thickness measurement methods to coatings Substrate

Coating/Method(a) Copper Nickel

Magnetic steel C, M (including corrosion-resistant steel) C, E(e) Nonmagnetic stainless steels Copper and alloys C, only on brass and Cu-Be C Zinc and alloys Aluminum and B, C alloys Magnesium and B alloys C Nickel B Silver Glass sealing M nickel-cobalt-iron alloys (UNS K94610) B, C, E(e) Nonmetals

Chromium Electroless nickel

Zinc Cadmium Gold Palladium Rhodium Silver

Tin

Lead

Tinlead alloys

C, M(b) C, M

C(c), M(b)

C, M

B, C, M

B, M B, M

B, M

B, C, B, M C, M

B, C, B(d), M C(d), M

C, M(b) C

C(c)

C

B, C

B

B

B

B, C

C, M(b) C

C(c)

C

B, C

B

B

B

B, C, B, E(e) C B, C B, C

M(b) … B, C, B, C M(b) B, M(b) B

… B, E(b)(c) B

B B, C

B B

B B

B B

B B, C

B

B

B

B

B

B

…

…

C

B, C

B

B

B

B, C

B C(c), M(b)

B M

… B, M

B … B, M B, M

… B, M

… B, M

B

B

B

B, C

B

B

B

B

C

B, M(b) B C, M(b) M

… C(c), B, C

B, C, B, C B, C(c) B, C B, C (b) M B B, M(b) … B, E(b)(c) B B Titanium (a) B, β backscatter; C, coulometric; E, eddy current; M, magnetic. (b) Method is sensitive to permeability variations of the coating. (c) Method is sensitive to variations in phosphorus content of the coating. (d) Method is sensitive to alloy composition. (e) Method is sensitive to conductivity variations of the coating. Source: Ref 17

B, C

B B, C B

B B, C

B, C … B, M

B, C

B, C B

B, C

B

B(d), C(d) B(d), C(d) B(d) B(d), C(d) B(d)

B(d), C(d) B, C B(d) B, C, B(d), M C(d), M

B

B(d), C(d) B(d)

Nonmetals

B, M

Vitreous and porcelain enamels M

B, E

E

B, E

E

B, E E

… E

E

…

B, E

…

B, E B, M

E …

…

…

B, E

…

Porosity is another coating property that is very sensitive to deposition parameters and to substrate preparation. For very thin gold coatings (less than about 1 μm, or 0.04 mil), substrate texture controls coating porosity. At greater thicknesses, porosity is controlled by the coating properties. The variation in porosity with coating thickness is shown in Fig. 6, which shows three distinct phases: substrate dominated, transition, and coating dominated. Porosity can also be controlled to a large extent through deposition parameters, and this effect is attributed to a reduction in deposit grain size. The effects of porosity with thickness for pulsed-plated gold are shown in Fig. 7; again, the same three phases are apparent, although they are displaced downward.

Fig. 6 Porosity versus deposit thickness for electrodeposited, unbrightened gold on a copper substrate. Compare with Fig. 7.

Fig. 7 Porosity versus deposit thickness for pulse-plated gold on a copper substrate. The curve for an unbrightened gold deposit on a copper substrate (top) is shown for comparison (see also Fig. 6). Electrodeposited coatings applied to thicknesses of less than about 25 μm (1 mil) are porous. Porosity decreases with thicknesses, but it is present even at greater thicknesses. One of the reasons for this is the codeposition of hydrogen at cathodes, owing to the fact that most plating processes operate at less than 100% cathode efficiency. With sacrificially protective coatings, such as zinc and cadmium on steel, porosity is of little significance. Within limits, substrate steel exposed in the pores will be cathodic to the zinc or cadmium coating. Thus, the coatings will corrode under adverse conditions to protect the exposed cathodic steel, at least until the coating has been consumed by conversion to corrosion products. With the exception of chromium, coatings that are cathodic to the substrate depend on freedom from porosity for protection of the less noble substrate. Accordingly, such coatings must be tested for porosity as well as for thickness. Examples of ASTM standards for porosity testing are given in Ref 16 and 18, 19, 20, 21. Adhesion is very important in the electrodeposition of a metallic coating, as implied in the definition of electroplating given earlier in this article. Therefore, it is logical for many end- product specifications for plated components to contain adhesion tests for verification of this requirement. ASTM B 571 (Ref 22) covers the following adhesion tests for metallic coatings: the burnishing tests, the chisel-knife test, the draw test, the file test, the grind-saw test, and the heat- quench test.

Corrosion. Testing for corrosion resistance is an important part of evaluating electroplated coatings and allied finishing techniques. Standard methods of corrosion testing are given in Ref 16 and 23, 24, 25, 26, 27. Accelerated tests of corrosion resistance of electroplated coatings can often lead to misleading conclusions. Claims that a certain number of hours in a particular test is equivalent to so many months or years of actual service should be viewed with suspicion.

References cited in this section 2. C.L. Faust, “Electroplating,” Course 22, Metals Engineering Institute, American Society for Metals, 1983 16. “Standard Method of Salt Spray (Fog) Testing,” B 117, Annual Book of ASTM Standards, American Society for Testing and Materials 17. “Thickness of Metallic and Inorganic Coatings,” B 659, Annual Book of ASTM Standards, American Society for Testing and Materials 18. “Standard Specification for Electroplated Engineering Nickel Coatings,” B 689, Annual Book of ASTM Standards, American Society for Testing and Materials 19. “Standard Specification for Electrodeposited Coatings of Tin-Nickel Alloy,” B 605, Annual Book of ASTM Standards, American Society for Testing and Materials 20. “Standard Test Methods for Porosity in Gold Coatings in Metal Substrates,” B 583, Annual Book of ASTM Standards, American Society for Testing and Materials 21. “Standard Specification for Electrodeposited Coatings of Tin,” B 545, Annual Book of ASTM Standards, American Society for Testing and Materials 22. “Standard Test Methods for Adhesion of Metallic Coatings,” B 571, Annual Book of ASTM Standards, American Society for Testing and Materials 23. “Standard Recommended Practice for Rating of Electroplated Panels Subjected to Atmospheric Exposure,” B 537, Annual Book of ASTM Standards, American Society for Testing and Materials 24. “Standard Method of Testing FACT (Ford Anodized Aluminum Corrosion Test),” B 538, Annual Book of ASTM Standards, American Society for Testing and Materials 25. “Standard Methods for Corrosion Testing of Decorative Chromium Electroplating by the Corrodkote Procedure,” B 380, Annual Book of ASTM Standards, American Society for Testing and Materials 26. “Standard Method for Copper-Accelerated Acetic Acid-Salt Spray (Fog) Testing (CASS Test),” B 368, Annual Book of ASTM Standards, American Society for Testing and Materials 27. “Standard Method of Acetic Acid-Salt Spray (Fog) Testing,” B 287, Annual Book of ASTM Standards, American Society for Testing and Materials


Applications for Electrodeposited Coatings Nickel is widely used as a corrosion-resistant coating and as an undercoat for subsequent coatings. It is not resistant to HNO3 or to environments containing chloride ions (Cl-). However, it corrodes slowly in environments that do not contain Cl- ions. Nickel is widely used by the automotive industry as an underlayer for microcracked chromium to protect steel. As previously discussed, nickel is usually plated as part of a multilayer coating system in which the electrode potential of each layer is different from those of other layers. The outer layer in such systems is often 130 mV more noble or cathodic than the layer beneath it. Nickel is also used over copper as an underlayer for gold to prevent rapid diffusion of the gold into the copper substrate. Nickel and its reaction products are toxic and a substantial percentage of the public suffer “nickel itch,” an allergic reaction to prolonged intimate contact with nickel metal. Nickel with organic additives to produce a very bright finish is widely used, although these coatings are much more brittle than the semibright nickel. Nickel-phosphorus can be electrolessly or electrolytically deposited as a metallic coating. In general, its corrosion performance exceeds that of nickel in all environments. If the phosphorus content is above about 10 wt%, the coating is amorphous and therefore lacks grain boundaries or other crystalline defects at which corrosion can initiate. Electroless nickel has some advantage over electrolytic nickel in that its thickness and phosphorous content are uniform over the substrate and independent of substrate geometry. Electroless nickel is not resistant to hot caustic solutions. Heat treating nickel-phosphorus coatings above 350 to 400 °C (660 to 750 °F) results in recrystallization and loss of some corrosion resistance even though other properties, such as wear or hardness, may be enhanced. Diamond dust, Teflon globules, and other discrete particles are routinely plated into the nickel-phosphorus matrix to deliver wear characteristics not achievable from any other finish. Cadmium and zinc are compared in the section “Idealized Coatings” in this article. Cadmium is generally preferred for the protection of steel in marine environments, and zinc is preferred in industrial environments. Cadmium tends to embrittle high-strength steel somewhat less than zinc does (being process related, new deposition processes may change this). The coefficient of friction of cadmium is less than that of zinc; therefore, cadmium is preferred for fastening hardware and connectors that have to be taken on and off repeatedly. Cadmium is much more toxic than zinc, and applications in which its corrosion products can get into the environment must be avoided. There is worldwide movement toward restricting cadmium plating to critical applications or banning its use outright. Cadmium should be specified only after exhaustive investigation of the implications and alternatives. The corrosion performance of both zinc and cadmium is greatly enhanced by chromate conversion coatings (see the article “Chromate and Chromate-Free Conversion Coatings” in this Volume). Chromium is very resistant to atmospheric corrosion, but is soluble in HCl or in alkaline (caustic) solutions. Its resistance to corrosion is thought to be due to the formation of an amorphous chromium oxide that acts as a passive film to protect the metal. The hardness, low coefficient of friction, and ability to retain oil of heavy chrome deposits has made them an indispensible finish in thousands of moving-part applications. Chromium is deliberately deposited for decorative and wear applications as a microcracked coating over nickel so that corrosion currents are uniformly distributed over a large area. Chromium may be deposited as a noncracked coating both from hexavalent chromium (Cr6+) and from trivalent chromium (Cr3+) electrolytes. Tin is widely used as a barrier and as an anodic coating for steel and copper. Because it is readily solderable and its oxide is conducting, it finds application in electrical conductors and contacts. Its corrosion products are not toxic; therefore, tinned containers are widely used in the food industry. Recommended thicknesses for application over copper components are given in Table 4 and over ferrous components in Table 5.

Table 4 Suggested thicknesses for electrodeposited tin coatings on copper alloys containing a minimum of 50% Cu. Minimum local thickness μm mils 1.2 Contact with food or water where a complete cover of tin must be maintained 30 against corrosion and abrasion 15 0.6 Protection in atmosphere and in less aggressive immersion conditions 5 0.2 To provide solderability and protection in mild atmospheric conditions (a) 2.5–8 0.1–0.3(a) Coatings flow brightened by fusion (solderability and protection in mild atmospheres) (a) On brass, an undercoat of 2.5 μm (0.1 mil) thick, copper, nickel, or bronze is required. Source: Ref 28 Purpose of coating

Table 5 Suggested thicknesses for electrodeposited tin coatings on ferrous components Minimum local thickness μm mils 1.2 Contact with food or water where a complete cover of tin must be maintained 30 against corrosion and abrasion 20 0.8 Protection in atmosphere 0.4 Protection in moderate atmospheric conditions with only occasional condensation 10 of moisture 5 0.2 To provide solderability and protection in mild atmospheres 0.1–0.3 Coatings flow brightened by fusion (solderability and protection in mild 2.5–8 atmospheres) Source: Ref 28 Lead is stable in most atmospheres and in both chromic (H2CrO4) and sulfuric acid (H2SO4) electrolytes. However, it is affected by chlorides. Lead electrodeposits form very inert passive films that tend to reduce the effects of galvanic corrosion. Copper is not particularly corrosion resistant in the atmosphere, and it tarnishes rapidly. When used alone, it should be protected by an appropriate corrosion inhibitor, such as benzotriazole. On the other hand, copper is often present in coating systems because bright copper is leveling, and its small grain size tends to reduce the porosity and thus improve the corrosion resistance of subsequent coatings. Gold is often used decoratively and for electrical contact applications. Gold that is plated directly over copper will increase the corrosion rate of the copper through the unavoidable porosity of the gold deposit. Gold will also rapidly diffuse into copper. Therefore, electroplated nickel or cobalt barrier coatings are normally plated underneath the gold coating. For wear applications, alloys with nickel or cobalt are preferred; however, pure gold has a lower contact resistance. Pulse plating will reduce porosity and increase the hardness of electrodeposited gold. Purpose of coating

Reference cited in this section 28. L.L. Shreir, Ed., Corrosion Control, Vol 2, Corrosion, Newnes-Butterworths, 1976


Selection of Coatings Tables 6 and 7 represent current practice in guiding designers and platers to minimum thickness of the two principal classes of protective and decorative/protective coatings used in product finishing: zinc and cadmium coatings and bright nickel-chromium. Specification guides for these coatings are given in Ref 1 and are also contained in Federal and Military Specifications (Ref 6, 29). Frequently cited specifications in these documents are: Specification Title “Standard Specification for Electrodeposited Coatings of Cadmium on Steel” ASTM A 165 SAE-AMS-QQ-P- “Plating, Cadmium Electrodeposited” (Replaces MIL-QQ-P-416, cancelled 28 March 2002) 416 “Standard Specification for Electrodeposited Coatings of Zinc on Iron and Steel” ASTM B 633 “Standard Specification for Electrodeposited Coatings of Copper plus Nickel plus ASTM B 456 Chromium and Nickel plus Chromium” SAE-AMS-QQ-C- “Chromium Plating (Electrodeposited)” (Replaces MIL-QQ-C-320, cancelled 3 May 2001) 320 Anodic Coatings for Aluminum and Aluminum Alloys MIL-A-8625 International Standards Organization (ISO) standards often govern international transactions involving the buying and selling of plated parts. These standards are available from the American National Standards Institute (ANSI) or from ISO. Table 6 Recommended minimum thicknesses and typical applications for zinc and cadmium coatings electrodeposited on iron and steel Service conditions

Coating thickness(a) μm mils

Chromate finish

Time to white corrosion in salt spray, h

Typical applications

5

0.2

8

0.3

Severe (exposure to condensation; infrequent wetting by rain; and cleaners)

13

0.5

… 12–24 24–72 72–100 … 12–14 24–72 72–100 … 12–24 24–72 72–100

Screws, nuts and bolts, buttons, wire goods, fasteners

Moderate (mostly dry indoor atmosphere; occasional condensation, wear, and abrasion)

None Clear Iridescent Olive drab None Clear Iridescent Olive drab None Clear Iridescent Olive drab

Very severe (exposure to bold atmospheric conditions; frequent

25

1

None

…

Electrodeposited zinc Mild (indoor atmosphere; minimum wear and abrasion)

Tools, zipper pulls, shelves, machine parts

Tubular furniture, window screens, window fittings, builders' hardware, military hardware, appliance parts, bicycle parts Plumbing fixtures, pole line hardware

exposure to moisture, cleaners, and saline solutions; likely damage by abrasion or wear) Electrodeposited cadmium 5 Mild (see above)

0.2

Springs, lock washers, None … fasteners, tools, electronic Clear 12–24 and electrical components Iridescent 24–72 Olive drab 72–100 8 0.3 None … Television and radio Moderate (see above) chassis, threaded parts, Clear 12–24 screws, bolts, radio parts, Iridescent 24–72 instruments Olive drab 72–100 13 0.5 None … Appliance parts, military Severe (see above) hardware, electronic parts Clear 12–24 for tropical service Iridescent 24–72 Olive drab 72–100 (b) 25 1 None … … Very severe (see above) Clear 24 Iridescent 24–72 Olive drab 72–100 (a) Thickness specified is after chromate conversion coating, if used. (b) There are some applications for cadmium coatings in this environment; however, these are normally satisfied by hot dipped or sprayed coatings. Source: Ref 12 Table 7 Recommended minimum thicknesses and typical applications for electrodeposited bright nickel-chromium coatings Service conditions

Classifications(a)

Minimum thickness Nickel Chromium Nickel Chromium μm mils μm mils Steel, iron, or zinc die-cast substrates 4 10 0.4 0.1 0.004 Mild (normally warm, b dry, indoor p 4 10 0.4 0.1 0.004 atmospheres; minimum d 4 10 0.4 0.1 0.004 wear or abrasion) b mc 10 0.4 0.8 0.03 p mc 10 0.4 0.8 0.03 d mc 10 0.4 0.8 0.03 b mp 10 0.4 0.3 0.012 p mp 10 0.4 0.3 0.012 d mp 10 0.4 0.3 0.012 r 20 0.8 0.3 0.012 Moderate (indoor b exposure where p r 20 0.8 0.3 0.012 condensation may d r 20 0.8 0.3 0.12 occur, as in kitchens or b mc 15 0.6 0.8 0.03 bathrooms) p mc 15 0.6 0.8 0.03 d mc 15 0.6 0.8 0.03 b mp 15 0.6 0.3 0.012 p mp 15 0.6 0.3 0.012 d mp 15 0.6 0.3 0.012 r 30 1.2 0.3 0.012 Severe (occasional or d frequent wetting by rain d mc 20 0.8 0.8 0.03 or dew; possible d mp 25 1.0 0.3 0.012 exposure to cleaners p r 40 1.6 0.3 0.012

Typical applications

Household appliances, interior auto hardware, hair dryers, fans, inexpensive cooking utensils, coat and luggage racks, standing ashtrays, interior trash receptacles, inexpensive light fixtures

Steel and iron: stove tops, oven liners, home, office, and school furniture; bar stools, golf club shafts. Zinc alloys; bathroom accessories, cabinet hardware

Patio, porch, and lawn furniture; bicycles; scooters, wagons; hospital furniture; fixtures; cabinets

p mc 30 1.2 0.8 0.03 p mp 30 1.2 0.3 0.012 r 40 1.6 0.3 0.012 Auto bumpers, grilles, hubcaps, Very severe (damage d from wear or abrasion d mc 30 1.2 0.8 0.03 and lower body trim; light likely in addition to d mp 30 1.2 0.3 0.012 housings corrosive media) Copper and copper alloy substrates b r 5 0.2 0.13 0.005 Household appliances, oven Mild (see above) p r 5 0.2 0.13 0.005 doors and liners, interior auto d r 5 0.2 0.13 0.005 hardware, trim for major b mc 5 0.2 0.8 0.03 appliances, receptacles, light p mc 5 0.2 0.8 0.03 fixtures d mc 5 0.2 0.8 0.03 b mp 5 0.2 0.25 0.01 p mp 5 0.2 0.25 0.01 d mp 5 0.2 0.25 0.01 b r 15 0.6 0.25 0.01 Plumbing fixtures, bathroom Moderate (see above) hinges, light p r 15 0.6 0.25 0.01 accessories, fixtures, flashlights, and d r 15 0.6 0.25 0.01 b mc 10 0.4 0.8 0.03 spotlights p mc 10 0.4 0.8 0.03 d mc 10 0.4 0.8 0.03 b mp 10 0.4 0.25 0.01 p mp 10 0.4 0.25 0.01 d mp 10 0.4 0.25 0.01 d r 25 1.0 0.25 0.01 Patio, porch, and lawn furniture; Severe (see above) d mc 20 0.8 0.8 0.03 light fixtures; bicycle parts; d mp 20 0.8 0.3 0.012 hospital and laboratory fixtures p r 25 1.0 0.3 0.012 p mc 20 0.8 0.8 0.03 p mp 20 0.8 0.25 0.01 d r 30 1.2 0.25 0.01 d mc 25 1.0 0.8 0.03 d mp 25 1.0 0.25 0.01 r 30 1.2 0.25 0.01 Boat fittings, auto trim, hubcaps, Very severe (see above) d d mc 25 1.0 0.8 0.03 lower body trim d mp 25 1.0 0.25 0.01 (a) Nickel classifications: b, fully bright; p, dull or semibright; d, double-layer or triple-layer nickel coating, with the bottom layer containing less than 0.005% S and the top layer more than 0.04% S. If there are three layers, the middle layer should contain more sulfur than the top layer. Chromium classifications: r, regular (conventional) chromium; mc, microcracked chromium having more than 750 cracks/in.; mp, microporous chromium containing a minimum of 64,500 pores/in.2 that are visible to the unaided eye. Source: Ref 12 Nickel-Chromium Coatings. Because of their importance in the automotive industry, nickel-chromium (with or without a copper underlayer) coatings have been subjected to many changes and exhaustive testing over a long span of years. Results of extensive field experience and testing of variations dating from the days of buffed nickel coatings (1870 to 1924) are reported in Ref 30. Figure 8 shows a history of the development and use of nickel-chromium coatings. Significant developments in this history were the determination that the basic corrosion resistance of these systems is controlled by the thickness and composition of the nickel layer, the invention of the leveling copper and nickel plating processes, the development of semibright and bright nickel plating, the emergence of crackfree and microcracked chromium, and the evolution of duplex and triplex (multilayer) nickel-chromium systems. and saline solutions)

Fig. 8 History of the development of nickel and nickel-chromium coatings showing time of introduction and periods of use. Source: Ref 30 Reference 30 also discusses the application of these coatings on steel, zinc, plastics, and aluminum. In these systems, the chromium layer is very thin—of the order of only 0.25 μm (0.01 mils). However, the corrosion resistance of chromium plate depends not so much on its thickness as on its physical state. Crackfree, or nonporous, chromium is excellent, but does not remain crackfree in service. A small number of cracks are detrimental, but the presence of many fine microcracks may be beneficial (Ref 2). Microcracked chromium deposits cause the galvanic- corrosion action to be spread over a very wide area and therefore do not suffer from localized corrosion. Microcracked chromium, in combination with duplex or triplex nickel, is the most durable of these coatings. The results of marine exposure for 36 months on variations of these coatings on steel are shown in Fig. 9. Figure 10 shows results of a similar program for Ni-Cr and Cu-Ni-Cr coatings on zinc.

Fig. 9 Performance in a marine atmosphere of various types of nickel-chromium and copper-nickelchromium coatings on flat (a) and contoured (b) steel panels. ASTM performance rating: 10, best; 0, worst. Test duration: 36 months. Source: Ref 30

Fig. 10 Performance in a marine atmosphere of various types of nickel-chromium and copper-nickelchromium coatings on flat (a) and contoured (b) zinc panels. ASTM performance rating: 10, best; 0 worst. Test duration: 36 months. Source: Ref 30

References cited in this section 1. “Standard Definitions of Terms Relating to Electroplating,” B 374, Annual Book of ASTM Standards, American Society for Testing and Materials 2. C.L. Faust, “Electroplating,” Course 22, Metals Engineering Institute, American Society for Metals, 1983 6. M. Murphy, Ed., Metal Finishing Guidebook & Directory, Elsevier Science, 2001 12. Quality Metal Finishing Guide, Metal Finishing Suppliers' Association 29. Department of Defense Index of Specifications and Standards (DoDISS), United States Navy, Naval Publications and Forms Center 30. G.A. DiBari, Decorative Electrodeposited Nickel-Chromium Coatings, Met. Finish., Vol 75 (No. 6, 7), June–July 1977


Plating of Mill Products A trend toward applying tin and zinc coatings by electrodeposition to sheet and strip in tandem with continuous pickling and cold-rolling lines developed just before and during World War II. Much earlier, processes were developed in Germany and the United States to electrogalvanize wire. Advantages claimed over hot dip galvanizing were control of thickness, the fact that the substrate is not affected by heat, and that the ductility and formability of the substrate is not limited by the iron-zinc intermetallic compound layer formed on hot dip galvanized steel. In hot dip galvanizing, the edges, especially of strip steel, tend to destroy wipes. Therefore, special care is required to secure a uniform smooth coating (Ref 31). Electrolytic tinning was commercially introduced in 1927 (Ref 31). Before that time, all of the tinplate on the market was manufactured by hot dip processes. Again, because of the amenability of electrodeposition to close thickness control and the development of highly efficient pretreating and plating processes, electrotinning has captured practically all of the food packing business. Various electrolytic processes have been used for producing tinplate, including the acid processes—sulfate, fluoborate, phenolsulfonic, and halide (fluoridechloride)—and the alkaline sodium stannate bath. In addition to electrogalvanizing and electrotinning, continuous plating or cold-rolled sheet and strip is performed by integrated steel mills as well as specially continuous strip producers and coil coaters. The specialty houses continuously plate nickel on steel for the battery (mostly alkaline dry cell) industry. They also plate alloy coatings such as nickel-zinc for protective/decorative uses and copper- and copper alloy-plated steel for postfabrication of decorative items. Copper-nickel-chromium on nonferrous metals and wide-strip steel is also available. Reference 32 lists typical examples of decorative/protective applications of continuously plated Cu-Ni-Cr plated steel strip. Among these are various components for household appliances, dispensers, handles, lamps, cover plates for switches and outlet boxes, and furniture trim. Functional applications for copper-plated strip include flashlight shells and alligator clips for jumper cables. Nickel-plated strip is used in battery cans and caps and for paintbrush ferrules. There are few companies in the continuous strip plating field because of the high initial investment required and the specialized knowledge needed to operate this process successfully. Coil coaters, the name given to manufacturers of prepainted sheet and strip (mostly steel and aluminum substrates), apply thin zinc coatings plus chemical conversion coatings before applying organic coatings. Products of this segment of the industry are used in postforming operations. Excellent control of plating conditions and the absence of complex substrate geometries provide a high degree of uniformity in coatings on preplated strip. This is reflected in predictable corrosion resistance because uniform thickness leads to uniform corrosion resistance. Chromium plating has also penetrated the tin mill product field in such applications as food can ends. Thin coatings of tin and chromium act as a base for can varnishes and lacquers and for pigmented outer coatings. On the interior of cans, tin or chromium coatings are claimed to prevent filiform corrosion under the clear organic coatings—an insidious type of underfilm corrosion characteristic of clear and sparsely pigmented organic coatings. Information on equipment and processing is available in Ref 4, 31, and 33; Ref 33 also discusses corrosion of tinplate. Immersion Plating. Immersion, or chemical, deposition has a long history of limited industrial use—for example, tin immersion deposits on mill wire products (called liquor finishing), immersion nickel plating of steel before porcelain enameling and many types of ceramic coatings, and the immersion copper plating of wire to be used in inert-gas shielded arc welding. In the last application mentioned, pure copper and copper- tin (gold-colored) coatings have both been used as aids to wet drawing of the wire to finished size and to furnish a highly conductive surface in the welding gun. Both tin and copper have been used in wire mills for the wire used for such products as paper clips and staples. According to the corrosion theory, these coatings, which are

much more noble than the substrate, would be expected to hasten corrosion of the substrate. However, this is not the case, probably because the porosity is so widely distributed that whatever current is available for corrosion is so spread out as to be ineffective. This is similar to the effect of microdiscontinuous chromium (MDC).

References cited in this section 4. L.J. Durney, Ed., Electroplating Engineering Handbook, 4th ed., Van Nostrand Reinhold, 1984 31. H.E. McGannon, Ed., The Making, Shaping and Treating of Steel, United States Steel Company, 1964 32. H. Wettwer, Continuous Plating of Copper, Nickel and Chromium on Wide Steel Strip, Plat. Surf. Finish., Vol 61 (No. 5), May 1974 33. C.H. Mathewson, Zinc, ACS Monograph Series, Reinhold, 1959


Electrochemical Predictions of Corrosion Performance Accelerated tests used for the prediction of corrosion performance for a particular application of a coating system have a certain risk associated with them. It is very difficult to design a test environment to simulate the actual environment, but there is often no alternative. A number of accelerated electrochemical tests are given in Ref 34, 35, 36, 37. Sample Preparation. The preparation of the sample to be tested requires a great deal of care for a number of reasons. On an actual part, the geometry is rarely designed so that the current density is uniform over the entire part. Because such coating properties as grain size and porosity are critically affected by local current density, the ability of a coating to protect the part would be expected to vary from one point on the part to another. When conducting accelerated tests on flat samples (usually required by the tests listed in Ref 34, 35, 36, 37), the current density used to apply the coating(s) must be typical of that used on the real part. Equally important, the surface finish of the specimen substrate should be comparable to that of the actual substrate. The above tests usually require that the surface finish be produced by wet grinding on 600-grit SiC paper. Electrodeposited coatings are sometimes used without additional surface finishing. Grinding and electropolishing will certainly affect test results. Test Selection. Very high localized currents can exist during these tests and may result in rapid removal of the coating. The desired information is usually the corrosion current icorr, which can be correlated with actual weight loss. The polarization resistance measurement (Ref 34) is probably the safest of the tests for this determination. The suitability of the test should be determined by correlating actual performance with test results whenever possible. Crevice corrosion under the gaskets used in the tests may affect results and should be considered.

References cited in this section 34. “Standard Practice for Conducting Potentiodynamic Polarization Measurements,” G 59, Annual Book of ASTM Standards, American Society for Testing and Materials

35. “Standard Practice for Conducting Cyclic Potentiodynamic Polarization Measurements for Localized Corrosion,” G 61, Annual Book of ASTM Standards, American Society for Testing and Materials 36. “Practice for Standard Reference Method for Making Potentiostatic and Potentiodynamic Anodic Polarization Measurement,” G 5, Annual Book of ASTM Standards, American Society for Testing and Materials 37. “Standard Recommended Practice for Conventions Applicable to Electrochemical Measurements in Corrosion Testing,” G 3, Annual Book of ASTM Standards, American Society for Testing and Materials


Future Trends Due to the pressure to get cadmium out of the environment, and the fact that simple zinc plating is not a satisfactory substitute in many instances, the application of such alloys as tin-zinc, zinc-cobalt, zinc-iron, and zinc-nickel is growing rapidly. Aluminum coatings, whether applied by proprietary electroplating processes or other methods, are additional cadmium substitutes. The desire to remove nickel plating from jewelry and other applications where it can cause sensitivity reactions is fueling the development of alloy replacements. Chromium plating, while not considered a threat to the environment in its metallic form, does pose a threat to health and the environment during its application. Add the potential chromium plating has for hydrogen embrittlement, and substitutes such as cobalt plating, HVOF coatings, and TiN coatings applied by PVD are making inroads into some of its former strongholds. The hexavalent chromate coatings traditionally used to forestall white rust on aluminum, cadmium, and zinc have effectively been banned from automotive use. While trivalent chromate conversion coatings are the usual replacement, when this substitute is not adequate for specific applications, the aluminum or zinc plating may be foregone and electroless nickel plating used instead. The continuing miniaturization of the electronics world is impelling changes both toward and away from electroplating. Zinc plating has been found to cause troublesome zinc whiskers that foul computer cleanrooms and to which hot dip galvanizing is not vulnerable. While not strictly a corrosion-resistant application, copper electroplating is replacing aluminizing in electronic circuits and is the technology of choice for micro electro mechanical systems (MEMS) fabrications. Nickel electroplating/electroforming continues to grow stronger with society's ever-increasing reliance on miniature precision optics and electronics.


Acknowledgment

This article has been adapted from the article by J. Mazia and D. Lashmore, Electroplated Coatings, Corrosion, Vol 13, ASM Handbook (formerly 9th ed., Metals Handbook), ASM International, 1987, p. 419–431.


References 1. “Standard Definitions of Terms Relating to Electroplating,” B 374, Annual Book of ASTM Standards, American Society for Testing and Materials 2. C.L. Faust, “Electroplating,” Course 22, Metals Engineering Institute, American Society for Metals, 1983 3. F.A. Lowenheim, Ed., Modern Electroplating, 3rd ed., John Wiley & Sons, 1974 4. L.J. Durney, Ed., Electroplating Engineering Handbook, 4th ed., Van Nostrand Reinhold, 1984 5. F.A. Lowenheim, Electroplating, McGraw- Hill, 1978 6. M. Murphy, Ed., Metal Finishing Guidebook & Directory, Elsevier Science, 2001 7. R.S. Treseder, Ed., The NACE Corrosion Engineer's Reference Book, National Association of Corrosion Engineers, 1980 8. G.T. Beilby, The Hard and Soft States in Metals, J. Inst. of Met., No. 2, 1912, p 149–185 9. R.G. Baker, “Corrosion of Electrodeposited Coatings,” Slide Lecture 46, American Electroplaters and Surface Finishers Society, 1981 10. “Standard Practice for Design of Articles to be Electroplated on Racks with Nickel,” B 507, Annual Book of ASTM Standards, American Society for Testing and Materials 11. J. Hyner, Design for Plating, Electroplating Engineering Handbook, L.J. Durney, Ed., Van Nostrand Reinhold, 1984, p 50 12. Quality Metal Finishing Guide, Metal Finishing Suppliers' Association 13. “Standard Specification for Electrodeposited Coatings of Copper Plus Nickel Plus Chromium and Nickel Plus Chromium,” B 456, Annual Book of ASTM Standards, American Society for Testing and Materials 14. S. Wernick and R. Pinner, The Surface Treatment and Finishing of Aluminum and Its Alloys, Robert Draper, 1972 15. “Method for Mechanical Hydrogen Embrittlement Testing of Plating Processes and Aircraft Maintenance Chemicals,” F 519, Annual Book of ASTM Standards, American Society for Testing and Materials

16. “Standard Method of Salt Spray (Fog) Testing,” B 117, Annual Book of ASTM Standards, American Society for Testing and Materials 17. “Thickness of Metallic and Inorganic Coatings,” B 659, Annual Book of ASTM Standards, American Society for Testing and Materials 18. “Standard Specification for Electroplated Engineering Nickel Coatings,” B 689, Annual Book of ASTM Standards, American Society for Testing and Materials 19. “Standard Specification for Electrodeposited Coatings of Tin-Nickel Alloy,” B 605, Annual Book of ASTM Standards, American Society for Testing and Materials 20. “Standard Test Methods for Porosity in Gold Coatings in Metal Substrates,” B 583, Annual Book of ASTM Standards, American Society for Testing and Materials 21. “Standard Specification for Electrodeposited Coatings of Tin,” B 545, Annual Book of ASTM Standards, American Society for Testing and Materials 22. “Standard Test Methods for Adhesion of Metallic Coatings,” B 571, Annual Book of ASTM Standards, American Society for Testing and Materials 23. “Standard Recommended Practice for Rating of Electroplated Panels Subjected to Atmospheric Exposure,” B 537, Annual Book of ASTM Standards, American Society for Testing and Materials 24. “Standard Method of Testing FACT (Ford Anodized Aluminum Corrosion Test),” B 538, Annual Book of ASTM Standards, American Society for Testing and Materials 25. “Standard Methods for Corrosion Testing of Decorative Chromium Electroplating by the Corrodkote Procedure,” B 380, Annual Book of ASTM Standards, American Society for Testing and Materials 26. “Standard Method for Copper-Accelerated Acetic Acid-Salt Spray (Fog) Testing (CASS Test),” B 368, Annual Book of ASTM Standards, American Society for Testing and Materials 27. “Standard Method of Acetic Acid-Salt Spray (Fog) Testing,” B 287, Annual Book of ASTM Standards, American Society for Testing and Materials 28. L.L. Shreir, Ed., Corrosion Control, Vol 2, Corrosion, Newnes-Butterworths, 1976 29. Department of Defense Index of Specifications and Standards (DoDISS), United States Navy, Naval Publications and Forms Center 30. G.A. DiBari, Decorative Electrodeposited Nickel-Chromium Coatings, Met. Finish., Vol 75 (No. 6, 7), June–July 1977 31. H.E. McGannon, Ed., The Making, Shaping and Treating of Steel, United States Steel Company, 1964 32. H. Wettwer, Continuous Plating of Copper, Nickel and Chromium on Wide Steel Strip, Plat. Surf. Finish., Vol 61 (No. 5), May 1974 33. C.H. Mathewson, Zinc, ACS Monograph Series, Reinhold, 1959 34. “Standard Practice for Conducting Potentiodynamic Polarization Measurements,” G 59, Annual Book of ASTM Standards, American Society for Testing and Materials

35. “Standard Practice for Conducting Cyclic Potentiodynamic Polarization Measurements for Localized Corrosion,” G 61, Annual Book of ASTM Standards, American Society for Testing and Materials 36. “Practice for Standard Reference Method for Making Potentiostatic and Potentiodynamic Anodic Polarization Measurement,” G 5, Annual Book of ASTM Standards, American Society for Testing and Materials 37. “Standard Recommended Practice for Conventions Applicable to Electrochemical Measurements in Corrosion Testing,” G 3, Annual Book of ASTM Standards, American Society for Testing and Materials


Selected References • • • • • • • • • •

A. Brenner, Electrodeposition of Alloys: Principle and Practice, Academic Press, 1963 W. Canning, The Canning Handbook on Plating, 23rd ed., E. & F.N. Spon Ltd., 1982 J.K. Dennis and T.E. Such, Nickel and Chromium Plating, ASM International, 1994 A. Holmes, Rapid Spot Testing of Metals, Alloys and Coatings, Metal Finishing Information Services, 2002 J.B. Kusner and A.S. Kushner, Water and Waste Control for the Plating Shop, 2nd ed., Gardner Publications, 1981 F.J. LaManna, Electroplating Course Manual/ Basic Practical Electroplating, American Electroplaters and Surface Finishers Society, 1986 G.O. Mallory and J.B. Hajdu, Electroless Plating: Fundamentals and Applications, Noyes Publications, 1990 E.A. Ollard, Installation and Maintenance in Electroplating Shops, Robert Draper Ltd., 1967 M. Rubinstein, Electrochemical Metallizing: Principles and Practice, Van Nostrand Reinhold, 1987 W.H. Safranek, The Properties of Electrodeposited Metals and Alloys, 2nd ed., American Electroplaters and Surface Finishers Society, 1986

R. Leonard, Continuous Hot Dip Coatings, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 786–793

Continuous Hot Dip Coatings Ralph W. Leonard, Coated Steels Technology, LLC

Introduction HOT-DIP COATING is a process that primarily refers to the application of a low melting point metal as a coating on steel wherein the material to be coated is immersed into a molten bath of the coating metal. Hot-dip coating may involve the coating of steel sheet, strip, or wire, or it may involve the application of a coating onto forms such as pipe and structural shapes. The coating of steel sheet is defined as continuous hot-dip coating while the coating of piping, structural shapes, and discrete parts such as nails, screws, bolts, and nuts is called batch coating. Batch galvanizing was, in fact, done many years before the continuous coating process was developed. The earliest record of the use of zinc to protect steel dates back to 1836 (Ref 1). In batch galvanizing, the steel is fabricated into a finished part, cleaned to remove residual oils, pickled to remove oxides from the surface, and then dipped into molten zinc to apply the coating. As a result of long experience, the excellent protection that a zinc coating provides to steel was well known long before the continuous galvanizing process was developed more than 50 years ago. See the article “Batch Process Hot Dip Galvanizing” also in this Section. By far, the most common continuous hot-dip coated product is galvanized steel sheet. Galvanized sheet has a coating that is essentially zinc metal. Its companion product, galvannealed sheet, is produced in a similar way except that the coating of zinc, while molten, is alloyed via a diffusion process with iron from the steel sheet. The resultant coating has approximately 9 to 10% iron. Since steel sheet is the most commonly manufactured hot-dip coated product today, most of the ensuing discussion focuses on steel sheet issues, but much of the information would apply to hot-dip coated steel wire as well. The continuous hot-dip process is called “continuous” simply because the product intended to be coated, steel sheet for example, is passed in a continuous manner into a bath of the molten coating metal. As applied to steel sheet, the leading end of one coil of steel is welded onto the trailing end of the coil ahead of it in the processing line. With the use of an accumulator, which stores up to perhaps 1000 ft or more of sheet ahead of the central part of the processing line, these coil ends can be joined without stopping the central processing section. As a result, the sheet being processed through the coating bath never stops; the process is truly “continuous.” The primary reason that the hot-dip process is restricted to the application of low melting point metals is that the product to be coated (steel sheet) needs to have some amount of tensile strength as it is pulled through the coating line. Low carbon steel sheet, the most common hot- dip coated product, cannot be heated to temperatures much above 870 °C (1600 °F) without literally tearing apart as it is being transported through a large processing line. Thus, the coating metal has to have a lower melting point than 870 °C (1600 °F). Today, the highest melting point metal being applied to steel sheet is aluminum, which has a melting temperature of 660 °C (1220 °F). Since the continuous sheet-coating process was developed over 50 years ago for the application of zinc (galvanizing) onto steel sheet, the process has been employed for the application of other metallic coatings including aluminum and several aluminum/zinc alloys. Each of these coating types has special attributes that make them more suitable than galvanized sheet for specific types of applications. These coatings and their specific attributes will be discussed in the subsequent sections.

Reference cited in this section 1. “Hot-Dip Galvanizing for Corrosion Protection, A Specifier's Guide,” American Galvanizers Association

R. Leonard, Continuous Hot Dip Coatings, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 786–793 Continuous Hot Dip Coatings Ralph W. Leonard, Coated Steels Technology, LLC

Basic Principles Steel sheet is passed as a continuous ribbon through a bath of molten metal at speeds up to approximately 200 m/min (650 ft/min). A simplified schematic drawing of the strip as it is passed through the coating bath is shown in Fig. 1. One of the requirements of the process is that, while in the molten metal bath, the steel strip must react with the molten metal to form an intermetallic bond (often referred to as the alloy layer) across the entire surface of the steel sheet. This bond is present as a thin distinct region between the coating metal and the steel sheet. It is necessary for two reasons. One is that without the intermetallic-alloy layer the coating would dewet from the steel when the sheet exits the coating bath, and this would lead to either uncoated areas (voids) or a very nonuniform coating thickness. A uniform coating thickness is very important to achieve good corrosion resistance. Second, without the alloy layer, the coating would not adhere to the sheet when it is subsequently formed into a shaped part.

Fig. 1 Basic arrangement of the coating bath in a continuous hot-dip coating operation Wetting is the term describing the interfacial tension between a solid and a liquid. Here, dewetting is used to define the absence of the diffusion reaction between the steel surface and the molten coating metal. It is very important that this reaction take place across the entire surface area of the steel sheet. The intermetallic-bond layer has several important features. The inherent nature of the alloy layer formed on all hot-dip coated products, whether the coating is zinc, aluminum or one of the aluminum/zinc alloys, is that it is composed of one or more intermetallic compounds. These can be zinc-iron, aluminum-iron, or aluminum- zinciron compounds. By its nature, the bond layer is hard and brittle. For galvanized sheet, the alloy layer is

approximately ten times harder than the zinc coating outer layer. Thus, in order to be able to form the sheet without loss of coating adhesion, the alloy bond thickness must be kept as thin as possible. As the strip emerges from the molten bath at high speeds, it drags out excess metal, much like when an object is pulled rapidly from a container of water. The amount of excess metal depends on the speed at which the sheet is moving, the viscosity of the coating metal, and a number of other variables such as the surface roughness of the sheet, the presence of oxides on the surface of the bath metal, and the temperature of the bath metal. Through the use of an air-wiping process, a controlled amount (thickness) of molten coating, usually expressed as weight of coating per unit area, is allowed to remain on the strip surface. The molten coating is cooled, and then, following several other processing steps that may be used to modify the surface roughness, flatten the steel, and apply a rust-inhibitive surface treatment, the steel is recoiled at the exit end of the processing facility. No matter which type of coating metal is applied, the common features of hot-dip coated sheet products are shown schematically in Fig. 2. There is an outer coating layer that has essentially the same composition as the bath metal. This is followed by the intermetallic bond layer, a layer that forms by diffusion between the coating bath and the steel substrate. Beneath the alloy layer is the steel substrate.

Fig. 2 Schematic drawing of the cross section of a hot-dip coated sheet product showing the three features of all hot-dip coated sheet products—the outer coating layer, the intermetallic alloy bond zone, and the steel sheet substrate Process Details. Most commonly, the hot-dip process for steel sheet involves the following sequential steps: 1. An entry-end welder joins the trailing edge of one coil to the leading edge of the succeeding coil to allow the process to be truly continuous. 2. An alkaline cleaning section removes the rolling oils, dirt, and iron fines (surface contaminants from the cold reduction process) that are on the sheet surface. 3. An annealing furnace heats the steel to high temperatures to impart the desired mechanical properties (strength and formability) to the steel sheet. 4. A bath of molten metal contains the metal being applied to the steel surface. 5. A cooling section cools the strip and solidifies the coating as it emerges from the coating bath. 6. A temper mill imparts the desired surface finish to the coated steel. 7. A tension leveler flattens the strip to meet the end-use requirements. 8. A treatment section applies a clear, water- base treatment to the coating to prevent storage stains that can form on the coating surface when moisture is present (as condensation and/or water infiltration originating from improper shipping or storage). 9. An oiling section is used most often to apply a rust-inhibitive oil. At times, it can be used also to apply a forming oil.

10. A recoiler rewinds the finished coil of steel. At this point, the finished coil of coated steel has: • • • •

The desired steel strength and formability An adherent, uniformly thick, corrosion resistant coating The desired surface finish and sheet flatness A clear chemical treatment and/or oil to help maintain the bright metallic finish that is a characteristic of hot-dip coated sheet products

Not all hot-dip coating lines include all the previously listed processing steps. For example, some do not include the aqueous cleaning stage, relying instead on flame cleaning in the entry end of the annealing furnace. Others might not have a temper mill; temper rolling is not necessary and perhaps not desired for some applications of hot-dip coated products. Also, there is a type of hot-dip coating process commonly called flux galvanizing that relies on using fluxes to apply a uniform and continuous coating. The final product is very similar to that obtained using the process described earlier, especially with respect to corrosion resistance. This process is not used much today, and it is used only for making galvanized sheet. None of the other aluminum or aluminum/zinc alloy coatings are applied using the flux process. Some references to the flux- coating process refer to it as the cold process in contrast with the term hot process used to define the lines that have in-line annealing furnaces. By far, the number of hot lines exceeds the number of cold lines. In North America, only several flux lines are operating as of 2003. Benefits of Coating Steel. It is well known that low-carbon steels are susceptible to corrosion when exposed to moisture. Not only is the corrosion aesthetically undesirable in most applications, but the corrosion rate can reduce the design life of the component. It is not unusual for steels to corrode at a rate of several mils of thickness per year (Ref 2). Not only does the surface take on the reddish color of iron oxide, but also, over time, the steel may lose a significant amount of its thickness, with a corresponding loss in mechanical properties. Steels can be painted to reduce the rate of corrosion and to make the product more attractive in appearance. However, the presence of paint alone does not ensure that steel is fully protected from corrosion. This is especially true at sheared edges or at scratches in the paint where rusting of the steel can cause paint blistering and continuing disbondment between the paint and the steel sheet (Fig. 3). Often, more protection is needed to reduce the corrosion rate of steel so that the aesthetic requirements are met and the strength and integrity of the steel sheet are not lost.

Fig. 3 Paint blistering and red rust staining at a scratch in painted cold-rolled sheet

Reference cited in this section 2. X.G. Zhang, Corrosion and Electrochemistry of Zinc, Plenum Press, 1996

R. Leonard, Continuous Hot Dip Coatings, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 786–793 Continuous Hot Dip Coatings Ralph W. Leonard, Coated Steels Technology, LLC

Hot-Dip Coatings The six principal types of coatings currently being applied by the hot-dip process are: • • • • • •

Galvanized (zinc or zinc with aluminum added) Galvannealed (zinc with 8 to 10% Fe) 55Al-Zn coating 95Zn-Al coating Aluminized with pure aluminum Aluminized with aluminum plus silicon

Additionally, a product called terne-coated sheet is produced. Terne-coated sheet has a coating of lead plus 3 to 15% tin. Years ago, terne- coated sheet was the primary product used to manufacture fuel tanks. Also, it was used for roofing applications. Today, these markets are quite small. Each of these coatings except the galvannealed sheet are produced using a specific coating- bath composition. Galvannealed sheet is not produced from a zinc and iron coating-bath composition. Instead, the coating bath is the same as that used to make galvanized sheet except that the aluminum level in the coating bath is reduced from approximately 0.17 to 0.20% to 0.14% or less. The iron is derived from the substrate. The 95Zn-Al coating may contain either a small addition of mischmetal (a combination of rare earth elements) or magnesium. Galvanized Sheet. It was discovered more than 200 years ago that zinc, in intimate contact with steel, will protect the steel. This protection takes place by two mechanisms; one is simple barrier protection wherein the zinc simply prevents contact between the steel and the environment of moisture and oxygen. The second is galvanic protection that protects steel even at exposed areas such as sheared edges of a sheet or scratches in the coating. In order to galvanically protect steel, a metallic coating such as zinc has to be electrochemically more active than the steel in the specific environment where the product is being used. Fortunately, zinc is indeed more chemically active than steel in almost all environments. In the chemical reaction that takes place when galvanic protection occurs, zinc becomes the anode of a corrosion cell and protects the steel from corroding by making the steel the cathode. Reference 2 contains a very detailed description of the fundamentals of the corrosion and electrochemistry of zinc. Galvanic protection can be accomplished whether the zinc is attached to the steel with a wire or is in intimate contact. Since corrosion is an electrochemical reaction, the zinc and steel need to be attached in a fashion that allows free electron flow between the steel and zinc. One of the classic examples of the galvanic protection obtained with the use of zinc is the protection of steel ship hulls from corrosion in sea water by selective placement of zinc anodes around the outside of a ship. When zinc is applied over the entire surface of a steel as in the hot-dip galvanizing process, the steel is fully protected from corrosion by the presence of the zinc coating. The intimate bond that occurs between the steel and the coating metal offers complete electrical contact so that free flow of electrons necessary for galvanic protection can occur. One needs to keep in mind that, as the zinc is protecting steel on a galvanized product, it undergoes corrosion. That is, the zinc itself corrodes and is consumed over time. The benefit afforded to steel is that, in most environments, zinc corrodes at a significantly lower rate than steel. Thus, a thin zinc coating can provide longterm protection to steel. The reason that zinc normally corrodes at a much lower rate than steel is that the products of corrosion that form on the surface of the zinc develop a protective, somewhat passive film. When steel corrodes in the environment, the iron oxide that forms as a result of the corrosion reaction is not usually retained on the surface

of the steel as a protective, passive film. Typically, it spalls off and continually exposes a fresh steel surface to the air and moisture. The protective, passive film that forms on zinc in most atmospheres is a zinc carbonate film (Ref 3). This carbonate film forms by a reaction between the zinc corrosion products (zinc hydroxide and zinc oxide) with carbon dioxide in air. This carbonate film is often stable and remains as a protective covering over the zinc coating. Typically, the zinc carbonate film does not completely protect the zinc from further corrosion; it simply slows down the corrosion reaction. The zinc surface exposed to the atmosphere is not completely passivated, so corrosion of the zinc can continue. This phenomenon is actually needed to continue to provide galvanic protection to steel. If the passive film were complete and corrosion of the zinc stopped completely, the galvanic protection to steel at exposed edges and scratches would be lost. Zinc carbonate films can be broken down (dissolved) by reaction with moisture in the environment, especially acid rain. Thus, even though zinc corrodes at a much lower rate than low carbon steel, it does corrode in most atmospheres, and the net result is that the zinc is eventually consumed. The actual rate of corrosion is highly dependent on the type of exposure. References 2 and 4 contain extensive information on the corrosion rates for zinc in many different types of environments. Table 1 (Ref 5) contains a summary of the relative corrosion rates for steel versus zinc in various types of atmospheric environments. These data provide general guidance only; more detailed information for specific environments is contained in Ref 2 and 4. Table 1 Typical corrosion rates for zinc and steel Environment classification Corrosion rate Zinc Steel μm/yr mil/yr μm/yr mil/yr 8).

In open recirculating waters, phosphate treatment levels should be 15 to 20 mg/L (ppm). At lower concentrations of a few mg/L (ppm), ions will cause pitting attack. The ability of phosphate to inhibit steel is primarily due to the oxygen content of the system (Ref 22). The dissolved oxygen produces a defective thin film of γFe2O3. The fills in the voids and accelerates film growth. This prevents further diffusion of 2+ Fe ions from the metal surface. Inhibition by ions is sensitive to electrolyte concentration; Cl- ions can promote pitting attack. Because the protective oxide film contains voids and other inclusions, Cl- ions are easily adsorbed to form soluble complexes. Hydrolysis of these complexes produces acid domains, which leads to localized acidic attack. Film breakdown is a function of the aggressive ion concentration, but film repair will depend on oxygen concentration. Inhibition by

levels and

ions is sensitive to water quality and pH. There is minimal temperature sensitivity. One

word of caution regarding ions as low-carbon steel corrosion inhibitors: with waters of high calcium hardness, the potential for deposit formation increases with hardness, phosphate level, pH, and temperature. These deposits will stimulate crevice (underdeposit) corrosion. The need for an inhibitor to prevent calcium deposits becomes exceedingly more important as cooling waters increase in cycles of concentration and pH.

Cathodic Inhibitors Cathodic inhibitors suppress corrosion by reducing the effectiveness of the cathodic process. They do not cause intense local attack and are generally considered to be safe inhibitors. Cathodic inhibitors are not as effective as anodic inhibitors in reducing corrosion. Precipitating inhibitors fall under the domain of a cathodic inhibitor. They produce insoluble films on the cathode under conditions of locally high pH and isolate the cathode from the solution. Calcium bicarbonate [Ca(HCO3)2] will react with the alkaline medium at the cathode to form CaCO3. The increase in alkalinity is due to the oxygen-reduction process, generating OH- ions: (Eq 39) At the correct pH, CaCO3 will precipitate to form a hard, smooth deposit that prevents oxygen from diffusing onto the metal surface. The tendency of water to form a CaCO3 deposit is given by the LSI value. Zinc ions are used to achieve a general reduction in corrosion by precipitating as zinc hydroxide [Zn(OH)2] at the locally elevated pH of the cathode. Zinc is usually combined with phosphonates, polyphosphates, or orthophosphate. The durability of Zn(OH)2 is tenuous at best. Although not normally used alone in cooling systems, zinc is synergistic when combined with other inhibitors as part of a multicomponent treatment program. Zinc causes rapid development of a protective film over the metal surface. Above pH 7.5, the solubility of zinc is rapidly reduced to a few tenths of 1 mg/L (ppm). Its solubility can be significantly enhanced when combined with phosphonates or with some of the polymers used for Ca3(PO4)2 control. Polyphosphates are widely used cathodic inhibitors. They have been used for more than 60 years and are among the most economical of all inhibitor treatments. Sodium salts of the polyphosphates are normally used for corrosion control. They exist as linear polymers having the general structure shown in Fig. 6. The lower members of the series having “x” equal to or less than 2 are crystalline in nature, that is, x = 0, orthophosphate; x = 1, pyrophosphate; x = 2, tripolyphosphate. Higher members of the series are glassy and have no definitive structure. Varying the ratio of nNa2O to mP2O5 will produce species of different chain lengths and physicochemical properties, even though the inhibitive action of these polyphosphates is not significantly altered. The pH of the environment directly affects the protective properties of the phosphates. Orthophosphate inhibits corrosion in a more alkaline environment than does pyrophosphate. The crystalline phosphates protect low-carbon steel at higher pHs than the glassy phosphates, which function better near neutral pH. Normal concentrations in recirculating waters are 15 to 20 mg/L (ppm). The pH should be maintained in the range of 6.5 to 7.5 if low-carbon steel and copper alloys are both part of the system metallurgy.

Fig. 6 General structure of sodium polyphosphate. x = 0, sodium phosphate; x = 1, sodium pyrophosphate; x = 2, sodium tripolyphosphate; x = 12 to 14, sodium polyphosphate The polyphosphates are relatively insensitive to electrolyte concentrations, but do require an increase in dosage level with an increase in water corrosivity. They are effective inhibitors in controlling galvanic attack between two dissimilar metals (Ref 23), but do not prevent deposition of cathodic species on the more active metal, that is, copper deposition on steel. Divalent metal ions, and Ca2+ in particular, are needed with the polyphosphates for effective inhibition of steel corrosion. The ratio of Ca2+ ion concentration to polyphosphate concentration should be at least 0.2, and preferably 0.5. The protective film develops through the formation of a positively charged colloidal complex that migrates to the cathode, forming an amorphous protective barrier. The cationic complex is accounted for by assuming that the calcium insets itself in the polyphosphate chains. The colloidal particles are the result of numerous cationic complexes loosely knit one to another through the Ca 2+ ions. Polyphosphates will slowly revert to ions, in the presence of Ca2+ ions and alkaline pH. The phosphonates differ from the polyphosphates by a direct bond formation between the phosphorus and carbon atoms, rather than an oxygen-phosphorus bond in the polyphosphate. This linkage creates greater hydrolytic stability, which reduces the problem of reversion common to the polyphosphates. Two classes of materials are extensively used: AMP and HEDP. Their structures are shown in Fig. 7 and 8. When used alone, both compounds are marginally good low- carbon steel inhibitors, especially at high pH (pH 8.0). They protect ferrous metals through a mixed mode of inhibition. Incipient local attack occurs, but is rapidly arrested. Normal treatment levels are in the range of 15 to 20 mg/L (ppm). The hydrolytic stability of the phosphonates eliminates the problem of Ca3(PO4)2 deposition in moderately to highly alkaline waters. However, AMP will ions and therefore should not be fed concurrently with degrade in the presence of chlorine, producing chlorine. The phosphonates form very stable complexes with a variety of nonferrous metal ions and thus accelerate their corrosion, especially the copper-base alloys. The addition of zinc significantly suppresses this antagonistic attack. Calcium phosphonate deposits will form in moderate-hardness high-alkaline waters, that is, 400 to 500 mg/L (ppm) calcium hardness at pH 8 to 8.5. These deposits can be controlled with some of the new polymers available for Ca3(PO4)2 control.

Fig. 7 Structure of nitrilotris(methylenephosphonic) acid, AMP

Fig. 8 Structure of 1-hydroxyethylidene-1,1-diphosphonic acid, HEDP

The phosphonates are sensitive to overall water quality. The protection afforded to low-carbon steel decreases with increasing aggressive ion concentration. Sensitivity to temperature is also exhibited. The major difference between AMP and HEDP is that the latter resists oxidation to a greater extent and can therefore be used in the presence of chlorine. Multicomponent Systems. Cooling water formulations containing mixtures of inhibitors usually offer increased protection to ferrous metals. Such mixtures are synergistic in their action. Many combinations have been developed to achieve enhanced protection under a wide variety of plant operating conditions. Heavy-metal formulations containing zinc and/or chromate have been widely used in industry, where environmentally acceptable, particularly for severe corrosion problems. Nonheavy-metal treatment programs (no zinc or chromium) are receiving increased attention because of government discharge restraints. Complex blends of , polyphosphate, phosphonate, copper corrosion inhibitor (see the section “Copper Inhibitors” in this article), and Ca3(PO4)2 dispersants are probably the most widely used multicomponent inhibitors. Zinc Polyphosphate. The addition of zinc to the polyphosphates does not appreciably change the general nature of the polyphosphates. This system retains its insensitivity to electrolyte concentration, its threshold inhibition of CaCO3 and CaSO4, its ability to protect ferrous and nonferrous metals, and its detergent properties. Zinc polyphosphate inhibitors also enable multivalent metal ions to form positively charged colloidal complexes. Zinc increases the rate at which the protective film is formed on the metal surfaces (Ref 24). This rapid protective film formation improves the general corrosion protection to the system. It is also synergistic in combination with polyphosphates. The amount of treatment needed using a zinc-polyphosphate inhibitor combination is less than that required for the polyphosphates alone. The mode of inhibition is cathodic, similar to the polyphosphates in calcium-containing waters. The zinc accelerates film formation, restraining attack until a thin, tenacious, durable film is developed. The deposition of zinc phosphate [Zn3(PO4)2] does not seem to be involved in the inhibition process. Approximately 10 to 20% of the combined zinc polyphosphate blend is zinc. Good corrosion control is achieved through the synergistic interaction of zinc with polyphosphate. Beyond 20% Zn, little improvement in low-carbon steel corrosion protection is observed. Maintenance concentrations are usually 10 mg/L (ppm) as polyphosphate. Good practice requires a dosage of two to three times the maintenance level as a pretreatment to the system for a short time, that is, less than 1 week. A pH range of 6.8 to 7.2 is required for good control. This limitation in pH is necessary to prevent excessive attack on copper-base alloys (the lower the pH, the greater the attack on copper). Sensitivity to bulk water temperatures is minimal. Zinc Phosphonates. The combination of zinc with phosphonates provides significantly improved protection compared to phosphonates alone. The addition of zinc makes this formulation synergistic in its protection to carbon steel. Good corrosion control can be attained using 20 to 80% of the combined zinc phosphonate blend as zinc, with 30 to 60% being optimum (Ref 25). Overall treatment levels are 8 to 10 mg/L (ppm) phosphonate with 2 to 3 mg/L (ppm) Zn2+ ions. The need for zinc becomes paramount in the presence of copper-base alloys. When used alone, the phosphonates are aggressive to copper and form a strong copper-phosphonate complex. Zinc negates this antagonistic effect by forming a stronger and more stable complex, significantly reducing attack on the alloys. The effectiveness of the zinc-phosphonate combination is due to increased cathodic protection. The zinc effectively counteracts the anodic character of the phosphonates through the formation of a phosphonate-zinc complex, which is cathodic compared to the phosphonates alone. Due to this complex, broad pH application becomes possible. The zinc phosphonate treatment can be used over a pH range of 6.5 to 9. Zinc is held in solution at these more alkaline pH levels, and protection actually improves with an increase in pH. Little sensitivity is shown with increased salt or electrolyte concentration, and temperature effects are minimal. Thus, the zinc phosphonate systems can be used over a wide range of water quality, at high bulk water temperatures, 70 to 75°C (160 to 170°F), and with pH levels up to 9.0. Degradation of phosphonates to ions by chlorine is significantly reduced with a phosphonate-zinc system, and the latter can be used in chlorine environments because of diminished chlorine demand. The zinc stabilizes the complex, preventing the deterioration of the carbon- phosphorus bond in the oxidizing environment. Nonheavy-Metal Systems. Many combinations of building blocks are currently being used in nonheavy-metal formulations. Some of the systems in use include combinations of the phosphonates, AMP/HEDP; polyphosphate/ phosphonate mixtures, polyphosphate/HEDP; polyphosphate or phosphonate/orthophosphate

(HEDP/ ). These systems are synergistic in protecting low-carbon steel. They combine the many advantages of the individual components, including: •

• •

Complex formation with calcium and numerous polyvalent metal cations required to form the protective film over the metal surface. The polyphosphates still form the colloidal species for cathodic protection, and the phosphonate serves as a cathodic polarizer Threshold inhibition of slightly soluble inorganic salts, such as CaCO3 and CaSO4 Detergent and dispersive action on surface deposits

Most nonheavy-metal treatments for carbon steel function best in more alkaline or elevated- pH systems. Waters are less corrosive, but more prone to form scale or deposit. Therefore, good scale control is absolutely necessary. These treatments are generally not as effective as heavy- metal formulations. However, because of the less corrosive nature of the recirculating cooling system, good overall protection can be achieved. More care and attention are required to operate these nonheavy-metal treatments because they are less tolerant of system upsets and function under more contained ranges of water corrosivity, pH, and temperature. The AMP/HEDP nonheavy-metal inhibitor pair is classified as a cathodic inhibitor. A critical AMP/HEDP ratio of 1.5 to 1 is needed for optimal carbon steel protection (Ref 26). Performance improves with pH and should be used in systems having a pH of at least 7.5. The combined phosphonate program is not as sensitive to temperature as the individual components, nor is it adversely affected by high levels of Cl- and ions. Normal treatment levels of at least 15 mg/L (ppm) total phosphonate can be used to achieve good corrosion protection, 0.04%) without carbidestabilizing elements (titanium and niobium). Examples are AISI types 302, 304, 309, 310, 316, 317, 430, and 446. Susceptible higher- nickel alloys include alloy 600 (UNS N06600), alloy 601 (UNS N06601), alloy 800 (UNS N08800) (despite the presence of titanium), alloy 800H (UNS N08810), alloy 200 (UNS N02200), alloy B (UNS N10001), and alloy C (UNS N10002). Intergranular corrosion in these alloys is avoided by one or more of the following: • • •

Keeping the alloy in the solution heat treated condition at all times Limiting interstitial elements, primarily carbon and nitrogen, to the lowest practical levels Adding carbide-stabilizing elements, such as titanium, niobium, and tantalum, along with a stabilizing heat treatment where necessary

In general, these alloys are purchased from the mill in the solution heat treated condition, a requirement of most specifications. With regard to the austenitic stainless steels, solution heat treating consists of heating to a minimum temperature of 1040 °C (1905 °F) to dissolve all the carbides, followed by rapid cooling in water or air to prevent sensitization. Preserving the solution-treated condition is difficult, except for certain applications where reheating is not a requirement of fabrication. One example is a pump shaft machined from forged or rolled and solution heat treated bar stock. However, most applications for these alloys, such as piping components and pressure vessels, require hot rolling (for example, of plate) or hot bending (for example, of pipe) and welding; these practices, as discussed earlier, can cause sensitization. Solution heat treating after fabrication is generally impractical because of the possibility of irreparable damage due to distortion or excessive scaling or because of the inability to cool rapidly enough through the critical sensitization temperature range. In most applications involving exposure to environments that cause intergranular corrosion, the lowcarbon/nitrogen or stabilized alloy grades are specified. The new ferritic stainless steel UNS S44800 with 0.025% C (maximum) plus nitrogen is replacing the higher-carbon 446 grade in applications in which resistance to intergranular corrosion is a requirement. Types 304L, 316L, and 317L, with carbon and nitrogen contents limited to 0.03% (maximum) and 0.10% (maximum), respectively, are used instead of their highercarbon counterparts. The newer nickel-chromium alloy 904L (UNS N08904), with 0.02% maximum carbon, and the high-nickel alloys C-22 (UNS N06022) and B-2 (UNS N10665), which have maximum carbon contents of 0.015% and 0.01%, respectively, are now being used almost exclusively in the as- welded condition without intergranular corrosion problems. These low carbon levels are readily and economically achieved with the advent of the argon oxygen decarburization (AOD) refining process used by most alloy producers. By limiting the interstitial element content, sensitization is limited or avoided entirely during subsequent welding and other reheating operations. However, designers should be aware of the fact that lowering the carbon/nitrogen content also lowers the maximum allowable design stresses, as noted in appropriate sections of applicable fabrication codes, such as ASME section 8, division 1 for unfired pressure vessels and ANSI B31.3 for process piping. Commercially pure alloy 200 is a special case. With a maximum carbon content of 0.15%, alloy 200 will precipitate elemental carbon or graphite in the grain boundaries when heated in the range of 315 to 760 °C (600 to 1400 °F). This results in embrittlement and susceptibility to intergranular corrosion in certain environments, such as high-temperature caustic. Where embrittlement and intergranular corrosion must be avoided, alloy 201 (UNS N02201), with a maximum carbon content of 0.02%, is specified. Titanium as a carbide-stabilizing element is used in several ferritic and austenitic stainless steels, including types 409, 439, 316Ti, and 321, as well as the higher-nickel alloy 825 (UNS N08825), at a minimum concentration of approximately five times the carbon plus nitrogen content. In the same way, niobium, generally with tantalum, is used in types 309Cb, 310Cb, and 347 austenitic stainless steels at a minimum combined concentration of approximately ten times the carbon content. The higher-nickel alloys 20Cb-3, alloy 625, and alloy G contain even higher concentrations of niobium—up to a maximum of approximately 4% in the case of alloy 625. In general, when stabilized alloys are heated in the sensitizing temperature range, chromium depletion at the grain boundaries does not occur, because the stabilizing elements have a greater affinity for carbon than does chromium.

Under certain conditions, however, stabilized alloys will sensitize, especially during multi- pass welding or cross welding. They are also susceptible to a highly localized form of intergranular corrosion known as knifeline attack, which occurs in the base metal at the weld fusion line. In some cases, these alloys are given stabilizing heat treatments after solution heat treatment for maximum resistance to intergranular corrosion in the as-welded condition. For example, type 321 stainless steel is stabilize annealed at 900 °C (1650 °F) for 2 h, and alloys 825 and 20Cb-3 at 940 °C (1725 °F) for 1 h, before fabrication to avoid sensitization and knife-line attack. So treated, type 321 may still be susceptible, because titanium has a tendency to form an oxide during welding. As a result, its role as a carbide stabilizer may be diminished. For this reason, type 321 is always welded with a niobium-stabilized weld filler metal, such as type 347 stainless. Some specialty alloys have low interstitial element content plus the addition of stabilizing elements for resistance to intergranular corrosion. These alloys include the higher-nickel alloy G-3 (UNS N06985), which contains 0.015% C (maximum) and niobium plus tantalum up to 0.5%, and the newer ferritic stainless steels (Ref 3, 22, 26, 27) listed as follows: UNS No. Common name Content, % Cr Ni Mo C (max) N (max) Other (max) … 26 … 1 0.01 0.015 0.2 Nb S44627 Nu Monit 25 4 4 0.025 0.035 0.80 Ti+Nb S44635 Sea Cure 26 2 3 0.03 0.04 1.0 Ti+Nb S44660 AL 29-4C 29 0.5 4 0.03 0.045 1.0 Ti+Nb S44735 The newer ferritic stainless steels were developed primarily for heat-exchanger tubing applications for use in place of the higher-carbon unstabilized type 446 stainless steel (UNS S44600). These ferritic stainless steels have many useful properties. As discussed previously, specific corrosion environments cause intergranular corrosion in specific metal and alloy systems. A wealth of information, both published (Ref 28) and unpublished, has been developed on corrodents that cause intergranular corrosion in sensitized austenitic stainless steels; a partial listing appears in Table 2. The low-carbon/nitrogen or stabilized grades are specified for applications in which the austenitic steels have satisfactory general and localized corrosion resistance in these environments and in which sensitization by such operations as welding and hot forming will undoubtedly occur. Table 2 Partial listing of environments known to cause intergranular corrosion in sensitized austenitic stainless steels Environment Nitric acid

Concentration, % 1 10 30 65 90 98 50–85 Lactic acid 30 Sulfuric acid 95 99.5 Acetic acid 90 Formic acid 10 (plus Fe3+) 10 Chromic acid 10 (plus Fe3+) Oxalic acid 60–85 Phosphoric acid 2 (plus Fe3+) Hydrofluoric acid 5 Ferric chloride 25 Acetic acid/anhydride mixture Unknown Unknown Maleic anhydride

Temperature, °C (°F) Boiling Boiling Boiling 60 (140) to boiling Room to boiling Room to boiling Boiling Room Room Boiling Boiling Boiling Boiling Boiling Boiling 77 (170) Boiling Room 100–110 (212–230) 60 (140)

Unknown 49 (120) Cornstarch slurry, pH 1.5 … Room Seawater 66–67 75 (167) Sugar liquor, pH 7 Unknown 232 (450) Phthalic anhydride (crude) Evaluation tests for intergranular corrosion are conducted to determine if purchased materials have the correct chemical composition and are in the properly heat treated condition to resist intergranular corrosion in service (Ref 29). Evaluation testing is imperative for a number of reasons: • • • •

The stainless steels and nickel-base alloys to which these tests are applied are relatively expensive. These materials are frequently specified for critical applications in the petrochemical, process, and power industries. The principal cost associated with a corrosion failure is generally that of production loss, not replacement. For maximum cost-effectiveness, these materials should be used in their best possible metallurgical and corrosion-resistant conditions.

Most of the evaluation tests are described in detail in ASTM standards (Ref 13, 15, 16, 17). Appropriate evaluation tests and acceptance criteria for wrought alloys are listed in Table 1 in the article “Evaluating Intergranular Corrosion” in this Volume. Intergranular corrosion is rare in nonsensitized ferritic and austenitic stainless steels and nickel- base alloys, but one environment known to be an exception is boiling HNO3 containing an oxidizing ion such as dichromate (Ref 30), vanadate, and/or cupric. Intergranular corrosion has also occurred in low-carbon, stabilized and/or properly solution heat treated alloys cast in resin sand molds (Ref 31). Carbon pickup on the surface of the castings from metal-resin reactions has resulted in severe intergranular corrosion in certain environments. Susceptibility goes undetected in the evaluation tests mentioned previously, because test samples obtained from castings generally have the carbon-rich layers removed. This problem is avoided by casting these alloys in ceramic noncarbonaceous molds. Other metals, such as magnesium, aluminum, lead, zinc, copper, and certain alloys, are susceptible to intergranular corrosion under very specific conditions. A copper pipe that ostensibly conveyed grade 2 fuel oil from an UST to a furnace displayed intergranular attack (Fig. 3); the mechanism is not properly understood. Very few case histories are reported in the literature.

Fig. 3 (a) Intergranular cracking in 9.5 mm ( in.) diameter copper pipe that conveyed No. 2 fuel oil from an underground storage tank. (b) Micrograph of intergranular crack. Dichromate etch. 63× An unusual form of intergranular corrosion known as exfoliation, which occurs in aluminum-copper alloys, is discussed in the section “Exfoliation” in this article. Intergranular SCC is discussed in the following section.

Stress-Corrosion Cracking Stress-corrosion cracking is a type of environmental cracking caused by the simultaneous action of a corrodent and sustained tensile stress. The following discussion deals primarily with anodic SCC. Anodic SCC is believed to be a delayed cracking phenomenon that occurs in normally ductile materials under the stress resulting from accelerated electrochemical corrosion at anodic sites of the material as well as at the crack tip. Other types of environmental cracking, such as hydrogen stress cracking and liquid metal embrittlement, are discussed later in this article. Table 3 condenses several sources of published data on corrodents known to cause SCC in various metal-alloy systems (Ref 6, 32, 33). These data cover the SCC environments of major importance to the materials engineer. This table, as well as other data published in the literature, should be used only as a guide for screening candidate materials for further in-depth investigation, testing, and evaluation.

Table 3 Some environment-alloy combinations known to result in stress-corrosion cracking Environment

Amines, aqueous Ammonia, anhydrous Ammonia, aqueous Bromine Carbonates, aqueous Carbon monoxide, carbon dioxide, water mixture Chlorides, aqueous Chlorides, concentrated, boiling Chlorides, dry, hot Chlorinated solvents Cyanides, aqueous, acidified Fluorides, aqueous Hydrochloric acid Hydrofluoric acid Hydroxides, aqueous Hydroxides, concentrated, hot Methanol plus halides Nitrates, aqueous Nitric acid, concentrated Nitric acid, fuming Nitrites, aqueous Nitrogen tetroxide Polythionic acids Steam Sulfides plus chlorides, aqueous Sulfurous acid Water, high-purity, hot

Alloy system Aluminum alloys … … … … … …

Carbon steels • • … … • •

Copper alloys • … • … … …

Nickel alloys … … … … … …

Stainless steels Austenitic Duplex … … … … … … … … … … … …

Titanium Martensitic alloys … … … … … … … … … … … …

Zirconium alloys … … … • … …

• …

… …

… …

• •

• •

• •

… …

… …

• …

… … … … … … … …

… … • … … … • …

… … … … … … … …

• … … • … • … •

… … … … … … • •

… … … … … … • •

… … … … … … • •

• • … … • … … …

… • … … … … … …

… … … … … … … … …

… • … … … … … … …

… • … … • … … • …

… … … … … … • … …

… … … … … … • … •

… … … … … … … … •

… • … … … … … … •

• … … • … • … … …

• … • … … … … … …

… •

… …

… …

… •

• …

… …

… …

… …

… …

Stress-corrosion cracking is not a certainty in the listed environments under all conditions. Metals and alloys that are indicated as being susceptible can give good service under specific conditions. For example, referring to Table 3: •

• • •

Anhydrous ammonia will cause SCC in carbon steels but rarely at temperatures below 0 °C (32 °F) and only when such impurities as air or oxygen are present; addition of a minimum of 0.2% H2O will inhibit SCC. Aqueous fluorides and hydrofluoric acid (HF) primarily affect alloy 400 (UNS N04400) in the nickel alloys system; other nickel alloys are resistant. Steam is known to cause SCC only in aluminum bronzes and silicon bronzes in the copper alloys system. Polythionic acid only cracks sensitized austenitic stainless steels and nickel alloys; SCC is avoided by solution annealing heat treatments or selection of stabilized or low-carbon alloys.

Stress-corrosion cracking is often sudden and unpredictable, occurring after as little as a few hours exposure or after months or even years of satisfactory service. Cracking occurs frequently in the absence of other forms of corrosion, such as general attack or crevice attack. Virtually all alloy systems are susceptible to SCC by a specific corrodent under a specific set of conditions, that is, concentration, temperature, stress level, and so on. Only the ferritic stainless steels as a class are resistant to many of the environments that cause SCC in other alloy systems, but they are susceptible to other forms of corrosion by some of these environments. SCC of Stainless Steels. The combination of aqueous chlorides and austenitic stainless steels is probably the most important from the standpoints of prevalence, economics, and investigation. Although the mechanism and boundary conditions for chloride SCC are still not fully defined, it is reasonably safe to state that chloride SCC of austenitic stainless steels: • • • • •

Seldom occurs at metal temperatures below 60 °C (140 °F) and above 200 °C (390 °F) Requires an aqueous environment containing dissolved air or oxygen or other oxidizing agent Occurs at very low tensile stress levels, such that stress-relieving heat treatments are seldom effective as a preventive measure Affects all the austenitic stainless steels approximately equally with regard to susceptibility, time to failure, and so on Is characterized by transgranular branchlike cracking, as seen under a metallurgical microscope

So many materials and environments are sources of chlorides that they are hard to avoid (Ref 32). Significant costs for repairs and replacements, as well as lost utility, have occurred through the years in the petrochemical industry as a result of chloride SCC: • • •

In water-cooled heat exchangers from chlorides in the cooling water Under thermal insulation allowed to deteriorate and become soaked with water that leached chlorides from the insulation Under chloride-bearing plastics, elastomers, and adhesives on tapes

In the case of shell and tube heat exchangers, in which chloride-bearing waters are used for cooling, a number of preventive measures are available. In vertical units with water on the shell side, cracking occurs most often at the external surfaces of the tubes under the top tubesheet (Ref 32). This is a dead space (air pocket) at which chlorides are allowed to concentrate by alternate wetting and drying of tubing surfaces. Adding vents to the top tubesheet sometimes alleviates this problem by eliminating the dead space and allowing complete water flooding of all tubing surfaces (Ref 32). See Fig. 4(f) in the article “Designing to Minimize Corrosion” in this Volume.

Fig. 4 (a) Stress-corrosion cracking of copper pipe under elastomeric insulation from an in-ground installation. (b) Micrograph of crack. Etched. 50× However, in many cases, this approach results in only a nominal increase in time to failure or no benefit whatsoever because of other inherent deficiencies, such as low water flows or throttling water flow to control process temperatures. Thus, a material more resistant to chloride SCC is required, such as the austenitic highernickel alloys 800 (UNS N08800) and 825 (UNS N08825); alloy 600 (UNS N06600); ferritic stainless steels such as type 430 (UNS S43000), 26Cr-1Mo (UNS S44627), and SC-1 (UNS S44660); duplex stainless steels such as alloys 2205 (UNS S31803), 255 (Ferralium 255, UNS S32550), and 2507 (UNS S32750); AISI type 329 (UNS S32900); or titanium. In this case, selection depends primarily on economics, requiring the least expensive material that will resist process-side corrosion as well as water-side SCC. Alloys are said to be either resistant or immune to chloride SCC, depending on how they perform in accelerated laboratory tests. In general, an alloy is immune if it passes the boiling 42% magnesium chloride test conducted according to ASTM G 36. Examples are alloy 600, 26Cr-1Mo, and grade 3 commercially pure titanium (UNS R50550). Industry has recognized the severity of this test and has devised other accelerated laboratory tests, such as boiling 25% sodium chloride and the Wick test (ASTM C 692), that are more representative of actual conditions in the field (Ref 34). Thus, alloys that fail the G 36 test but pass the C 692 test, such as alloy 800, alloy 2205 (UNS S31803), and 20 Mo6 (UNS N08026), typically provide many years of service as tubes in water-cooled heat exchangers. (The 300-series austenitic stainless steels, as a class, fail all the accelerated laboratory chloride SCC tests mentioned previously.) Unfortunately, all of the immune and resistant alloys are more expensive than most of the austenitic stainless steels. A more cost-effective application of some of these alloys involves a procedure known as safe ending, in

which short lengths (for example, 0.3 to 0.6 m, or 12 to 24 in.) are butt welded to the austenitic stainless steel tubes. The SCC-resistant ends are positioned in the exchanger at the point of greatest exposure to SCC conditions, that is, under the top tubesheet. Safe-ended tubes have extended the life of austenitic stainless steel tubing severalfold, with only a nominal increase in cost. Of course, the dissimilar metals must be weld compatible; this eliminates the use of titanium for safe ending. Another cost-effective answer to chloride SCC in heat exchangers is bimetallic tubing, in which the austenitic stainless steel required for process-side corrosion resistance is clad with a water-side SCC- and corrosionresistant material such as 90Cu-10Ni (UNS C70600) (Ref 32). Still another answer is cathodic protection with a sacrificial metal coating such as lead containing 2% Sn and 2% Sb applied by hot dipping or flame spraying (Ref 32). Chloride SCC under insulation can be prevented by keeping it dry, but this is easier said than done in many cases. It is particularly difficult in humid, high annual rainfall climates and where insulated equipment must be washed down periodically or is exposed to fire control deluge systems that are periodically activated (Ref 35). Under these conditions, three preventive measures are applied: •

•

•

Addition of sodium metasilicate as a SCC inhibitor to the insulation: At a minimum concentration of ten times the chloride content, the inhibitor is activated when the insulation becomes wet and is effective only when it wets the stainless steel surface. For maximum protection, metasilicate is painted on the vessel or piping before the installation of inhibited insulation. However, SCC has occurred after many years of service under inhibited insulation that was allowed to become so wet that the water-soluble inhibitor was leached out to a point below the minimum concentration required for prevention of SCC. Protective coating of the vessel or piping before insulation: Catalyzed high-build epoxy paints are effective to approximately 100 °C (212 °F), catalyzed coal tar/epoxy enamels to approximately 150 °C (300 °F), and silicone- base coatings to approximately 200 °C (390 °F). These coatings are even more effective if the stainless steel surfaces are heavily sandblasted before coating. Sandblasting peens the surface to a depth of 0.01 to 0.1 mm (0.4 to 4 mils), and this results in a layer under compressive stress that counteracts the tensile stresses required for SCC. Cathodic protection of the vessel or piping with aluminum foil under insulation (Ref 36): This method is claimed to provide both a physical barrier to chloride migration to stainless steel surfaces as well as cathodic protection when the insulation becomes wet, and it is effective at vessel temperatures between 60 and 500 °C (140 and 930 °F). However, the foil is attacked by the alkaline (generally) leachates from the insulation and must be renewed periodically.

Other important environments that cause SCC of stainless steels are hydroxide (caustic) solutions, sulfurous acid, and polythionic acids. Caustic SCC of austenitic stainless steels can be both transgranular and intergranular and is a function of solution concentration and temperature. It seldom occurs at temperatures below 120 °C (250 °F). At higher temperatures, the newer ferritic stainless steels, nickel, and high-nickel alloys provide outstanding service; 26Cr-1Mo stainless steel has found widespread application in heat-exchanger tube bundles serving caustic evaporators at 170 to 200 °C (340 to 390 °F). Nickel 200, nickel 201, and alloy 600 are resistant to 300 °C (570 °F) in caustic at concentrations up to 70% (Ref 11). Polythionic acid and sulfurous acid will cause SCC in sensitized nonstabilized austenitic stainless steels and nickel-base alloys. Cracking is always intergranular and requires relatively low tensile stresses for initiation and propagation. As-welded, normal carbon grades, such as types 304 and 316 and alloy 800, are particularly susceptible to SCC in weld HAZs. Low-carbon ( 30, the sampling distribution of the mean often approximates a normal distribution, even when the sampling universe or underlying population does not (Ref 5).

References cited in this section 4. W. Mendenhall and J.E. Reinmuth, Statistics for Management and Economics, Duxbury Press, 1974, p 192 5. N.T. Kottegoda and R. Rosso, Statistics, Probability, and Reliability for Civil and Environmental Engineers, McGraw-Hill Book Company, 1997, p 15–16, 222, 245–257

B.P. Jones, Statistics for the Corrosionist, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 972–979 Statistics for the Corrosionist Barnie P. Jones, Oregon Department of Transportation

Simple Statistical Tests A Simple Test Based on the Normal Curve. One commonly used test is based on the central limit theorem. It is called Student's t, Student being a pseudonym of its inventor, W.S. Gossett (Ref 6). Consider the following example: Specifications for an industrial coating call for a minimum application thickness of 20 mil and an average application thickness of 30 mil. A coated surface is tested for coating thickness at 60 random locations over the surface. The coating exceeds the minimum 20 mil in every case. However, it remains to be determined whether the average thickness is 30 mil. Note that, in this case, direction is irrelevant, and the coating fails whether thicker or thinner than the specified 30 mils. There are five key steps in carrying out a statistical test. These are (1) making assumptions, (2) selecting a test statistic, (3) defining the critical region, (4) computing the test statistic, and (5) making a decision. Assumptions. There are many assumptions required in setting up this kind of test. However, among the key assumptions is that the sample of 60 locations is random. Also, if n were less than 30, it would be necessary to assume that X, the coating thickness over the surface, is normally distributed. These assumptions, along with a variety of other fairly obvious assumptions that the situation is as it has been described, constitute what are sometimes referred to as model assumptions. A second category of assumptions are the hypotheses. In this case, the research hypothesis could be that the average application thickness is not 30 mils. The null hypothesis can be stated as μ = 30. Note that while all assumptions present the same threat to validity of the results, only the null hypothesis is tested.

For samples greater than 30, it is not as critical to make the model assumption that a variable is normally distributed because the central limit theorem implies that, for large samples, the sampling distribution of the mean from a nonnormal universe is often normal. Selecting a Test Statistic. The test statistic is Student's t: (Eq 12) where μ = 30, n = 60, and and s can be computed directly from the sample data using Eq 1 and 3. At least two considerations apply to the selection of this particular test statistic. First, coating thickness is measured using a ratio scale, which means that a statistic based on the mean is applicable. Second, the sample is relatively small, and Student's t is designed for small samples. Defining the Critical Region. The sampling distribution for Student's t depends on α and on degrees of freedom (n - 1). The limit of the critical region for Student's t at α = 0.05 and 59 degrees of freedom for a symmetric or two- tailed test is t > 2.000 or t < -2.000. A two- tailed test was specified, because the specification is nondirectional, that is, the coating can fail to meet the specification if the average thickness is either too thick or too thin. The sampling distribution for Student's t is actually a family of distributions that resemble the normal distribution but are somewhat flatter, the relative flatness depending on the degrees of freedom. For samples exceeding n = 120, the distribution for t becomes normal. Thus, for small or large sample applications, this statistic is generally referred to as the t test. Computing the Test Statistic. To compute the test statistic, first a sample mean ( ) of 29.5 mil and a standard deviation (s) of 1.9 mil must be computed. The test statistic t = (29.5 - 30)/ [(1.9)/ ] = −2.021. Making a Decision. Because -2.021 < -2.000, the decision is to reject the null hypothesis, supporting the conclusion that the work fails to meet the specification. Note, however, that if the preselected α level had been 0.01, the critical region would have been t > 2.660 or t < -2.660, and the null hypothesis would not have been rejected. The p value associated with t > 2.021 is 0.046, only marginally less than 0.05. This illustrates the advantage of reporting p values in contrast to the practice of testing against pre-selected α criteria. Other Issues Related to Student's t. Note the similarity between the definition of the standard error in Eq 7 and the denominator of Eq 12 defining Student's t. Note that using the estimate s instead of the parameter σ to estimate the standard error requires the quantity n - 1 to replace N in the denominator. A brief, clear explanation for this adjustment can be found in Ref 5. A longer but more intuitively appealing explanation can be obtained from Ref 7. Two-Sample Tests. Note that in the preceding illustrations, μ is given or assumed. Consider a different situation, in which a series of test specimens is compared to a series of comparable control specimens. This simple experiment suggests the research hypothesis μ1 ≠ μ2 and the null hypothesis, H0, μ1 = μ2. This is referred to as a two-sample test. The statistical inference process is parallel to that described previously, except that instead of having one set of sample values and one assumed parameter, there are two sets of sample values. In general, the test seeks to determine whether the two sets of specimens could have been drawn randomly from the same universe. The test statistic, based on the difference between the samples, is: (Eq 13) The simplification is possible, because the null hypothesis assumes μ1 = μ2. Note that what is being implied by this expression is not a sampling distribution of means but a sampling distribution of differences. The denominator for the expression in Eq 13 represents the standard error of that sampling distribution. Computing the test statistic is straightforward, except that there are two ways to estimate the standard error: a pooled variance and a separate variance estimate. If it can be assumed that σ1 = σ2, then the pooled variance estimate can be obtained as: (Eq 14) However, if it is assumed that σ1 ≠ σ2, then the separate variance estimate can be obtained as:

(Eq 15) It should be noted that the test using the separate variance estimate is slightly less efficient than the test using the pooled variance estimate. It also should be noted that in cases where both samples are approximately equal, degrees of freedom = n1 + n2 - 2, but for unequal samples, there is a complex expression that is used to find degrees of freedom (Ref 7).

References cited in this section 5. N.T. Kottegoda and R. Rosso, Statistics, Probability, and Reliability for Civil and Environmental Engineers, McGraw-Hill Book Company, 1997, p 15–16, 222, 245–257 6. D. Salsburg, The Lady Tasting Tea: How Statistics Revolutionized Science in the Twentieth Century, W.H. Freeman and Company, 2001, p 25–32 7. H.M. Blalock, Social Statistics, revised 2nd ed., McGraw-Hill Book Company, 1979, p 205–206, 208– 218, 231, 299–325, 340, 348, 352–369, 388, 398, 554–551


Confidence Limits Sometimes, the goal of a statistical analysis is not to test a hypothesis but to evaluate an estimate by establishing upper and lower bounds. These bounds are called confidence limits or confidence intervals. The process of establishing confidence limits shares much with statistical hypothesis testing, but there are some key logical and procedural differences. Consider an example. Borings from a reinforced concrete structure are analyzed for chloride concentrations at different depths. The researcher is comfortable assuming that true chloride penetration is uniform at different surface coordinates over the zone of interest (at least for the sake of this illustration), but that there is measurement error in the data collection procedure. Consequently, borings are taken at 26 random locations. Specimens from each boring are captured at 6 mm (0.236 in.) increments to a depth of 100 mm (3.94 in.). The purpose of this investigation is to assess corrosion risk to the structure. More specifically, the goal is to determine whether chloride concentration is high enough to initiate corrosion at a 51 mm (2 in.) depth, the minimum cover for the reinforcing steel. In the sample, the average ( ) chloride concentration at a depth of 48 to 54 mm (1.89 to 2.13 in.) is 0.016%, and the standard deviation is 0.011%. The first step in establishing a confidence interval is determining a confidence level, which is analogous to selection of a critical region in statistical hypothesis testing—that is, the risk one is willing to take of making an error. For example, a 95% confidence level corresponds to a critical region of α = 0.05. In other words, with a 95% confidence level, the confidence limit is expected to contain the parameter in 95 and not to include it in the remaining 5 out of 100 samples. Given a confidence level of 95%, the researcher can proceed to construct a confidence interval around . To do so, one works backward, starting with a table of areas under the normal curve, or in this case, because n = 26, from a table for the distribution of t. For 95% confidence (the same as α = 0.05 for a two-tailed test) and n - 1 or 25 degrees of freedom, the table shows a value of 2.060. The confidence interval can now be constructed as:

(Eq 16) In other words, at a 95% confidence level, the true chloride concentration at a 48 to 54 mm (1.89 to 2.13 in.) depth is believed to be between 0.0114 and 0.0205%. Assuming the corrosion threshold is 0.02%, these results suggest that the structure could be at immediate risk of corrosion damage. For a more complete discussion of confidence intervals, see Ref 5 or 7.

References cited in this section 5. N.T. Kottegoda and R. Rosso, Statistics, Probability, and Reliability for Civil and Environmental Engineers, McGraw-Hill Book Company, 1997, p 15–16, 222, 245–257 7. H.M. Blalock, Social Statistics, revised 2nd ed., McGraw-Hill Book Company, 1979, p 205–206, 208– 218, 231, 299–325, 340, 348, 352–369, 388, 398, 554–551


Sampling Issues In all of the applications of statistical inference discussed so far, there has been an underlying assumption that observations were selected randomly from a universe. In the interests of efficiency in presentation, this assumption has not been belabored until now. In plain language, the basis of statistical inference is that if chance alone determines the selection of observations, then chance alone also determines observed differences in the attributes of those observations. In practice, there are a number of ways to obtain what is known as a probability sample. The condition that defines a probability sample is that every element in a probability sample has a known probability of selection. For a discussion of the various approaches to probability sampling, see Ref 7. Sampling and Randomization. Sometimes the goal of sampling is to make an inference from a sample to the universe from which the sample was drawn. Probably, the most familiar example is election polling. However, the general notion of random sampling also applies to random assignment in the context of an experiment. The former relates to establishing external validity for an inference from a sample to a larger universe or population. The latter relates to internal validity by establishing the basis for the inference that treatment and control specimens in an experiment are initially equivalent. Randomization in the design of experiments may have limited applicability in corrosion science. Randomization is most likely to be of value in situations where relatively large numbers of specimens can be included in the experiment, and/or there is variability in the specimens and relative uncertainty about variables that could affect the outcome of the experiment. Randomization is particularly effective in controlling unidentified or unknown confounding factors. Determining Sample Size. The methods subsequently described focus on a sample drawn from a finite population but can readily be adapted to the context of randomization for an experiment. The determination of sample size is analogous to establishing a confidence interval in reverse. Consider Eq 17 and 18: (Eq 17)

(Eq 18) where B represents the bound on the error, and Z represents the standard score on the normal curve associated with the desired level of confidence. In election polling and other surveys, B is usually referred to as the margin for error. Note that given B, one can solve for n, if a suitable estimate can be found for the standard deviation (σ). There are a number of strategies to obtain a value for σ. One is to look for a comparable data set from another study. A second is to conduct a small pilot study. A third is to substitute a worst case or make a bald assumption. The point here is that, in order to make a precise estimate of sample size, it is desirable to know something about the distribution of the variable in question. To summarize, three pieces of information are needed to estimate sample size. The first is a tolerable error or an analogous goal for precision. The second is a confidence level to represent the acceptable risk of being wrong, typically 90, 95, or 99%. The third is an estimate for the standard deviation of the variable in question. Sampling from a Finite Universe. Special sampling considerations apply whenever a sample is being drawn from a finite universe, and the sample represents a nontrivial fraction of the whole universe. There is, in fact, a correction that applies to variance estimates based on finite populations, such that: (Eq 19) The quantity (N - n)/N is called the finite population correction (fpc). It should be obvious that the fpc becomes increasingly important as n/N increases. The fpc can be ignored whenever n/N does not exceed 5% (Ref 8). Another important implication that can be drawn from Eq 19 is that whenever a sample is drawn from a large universe, the absolute sample size is primarily what determines error, while the size of the universe is irrelevant for practical purposes. This result may be counterintuitive for many people, but it can easily be verified that, other things being equal, a sample of 500 has virtually the same confidence interval when drawn from a universe of 100,000,000 as when drawn from a universe of 50,000.

References cited in this section 7. H.M. Blalock, Social Statistics, revised 2nd ed., McGraw-Hill Book Company, 1979, p 205–206, 208– 218, 231, 299–325, 340, 348, 352–369, 388, 398, 554–551 8. W.G. Cochran, Sampling Techniques, 3rd ed., John Wiley and Sons, 1977, p 24–25


Analysis of Variance Analysis of variance is a specific technique developed in the early 1920s by the British statistician R.A. Fischer for application to agricultural experiments. The simplest case of analysis of variance involves one dependent variable, measured as an interval or ratio scale, and one independent variable, measured as a nominal or ordinal scale and having three or more categories or levels. In cases of an independent variable with two categories, analysis of variance reduces to a two-sample t test (see Eq 13). Analysis of variance is based on the idea that the variance of a dependent variable can be partitioned into different components. In the simple case of one independent variable, these components are within and between

groups. More specifically, from Eq 3 and multiplying both sides by n, a new quantity, ns2, called the sum of squares, is obtained: (Eq 20) In order to proceed further, it is necessary to introduce some refinements in notation. Let Xij represent the “ith” score in the “jth” category or subgroup. Subscript j ranges from 1 to k, and k is the number of categories of the independent variable. Subscript i ranges from 1 to n1 within category 1, from 1 to n2 in category 2, and from 1 to nk for category k, where n1, n2, and nk represent the number of cases within categories. Conditional, or within-category, means ·1, ·2, and ·k and the overall, or “grand,” mean ·· can also be described. Finally, to differentiate summation within and between categories, symbols ∑i and ∑j are used to represent summation over i and j, respectively. Partitioning of variance is illustrated in Eq 21:

(Eq 21)

or the total sum of squares (TSS) equals the within-groups, or residual, sum of squares (WSS) plus the between-groups sum of squares (BSS) (Ref 7). Note further that the components can be rewritten from Eq 20 as: (Eq 22)

(Eq 23)

(Eq 24) Note that BSS is found by subtracting WSS from TSS. Consider Eq 24. If all the ·j conditional means equal the grand mean ··, then BSS equals 0, and WSS equals TSS. Furthermore, the more the ·j differ, the larger the BSS and the smaller the WSS. Analysis of variance uses the ratio of BSS to WSS, adjusted for degrees of freedom, to compute a test statistic based on the F distribution: (Eq 25) Example: Analysis of Variance in Coating Bond Strength Tests. A test bridge deck is treated with three different coatings, and pull tests are conducted to measure bond strength, with 20 tests completed for each coating. The critical region is set at the customary α = 0.05. Analysis of variance is based on the F distribution. The distribution of F depends on degrees of freedom for both within- and between-groups estimates, and, in the case of (k - 1) = 1, F = t2 and t can, in fact, be regarded as a special case of F (Ref 7). Consequently, tables constructed to find the critical region for F require three dimensions: one for n - k degrees of freedom, one for k - 1 degrees of freedom, and one for α. Consequently, these tables are generally printed with each of several customary levels of α on a separate page. The lower limit of the critical region for F with (n - k) = 57, (k - 1) = 2, and α = 0.05 is approximately 3.17. Results show an overall average bond of ·· = 203.5 psi. However, for coating 1, the average is ·1 = 204.5 psi, while for coatings 2 and 3, bond strengths are ·2 = 228.7 psi and ·3 = 178.3 psi, respectively. One approach to the analysis of these results would be to deal with the samples in pairs and to run three separate t tests on all possible differences of means: μ1 = μ2, μ2 = μ3, and μ1 = μ3. Analysis of variance allows tests of the

form H0: μ1 = μ2 = μ3 = … μk. The research hypothesis is that at least one mean is different. Analysis of variance, by simultaneously comparing all three means, is more efficient, because it uses all 60 observations. Also, and more important, running three separate t tests inflates the probability of rejecting at least one of three null hypotheses, just as drawing an ace from a deck of cards is more likely in three attempts than in one. In fact, the probability of rejecting one of three hypotheses at α = 0.05 based on chance alone is 0.05 + 0.05 + 0.05 = 0.15. Table 4 summarizes the results of the one-way analysis of variance. Note that the quantities labeled “Mean square” are obtained by dividing the sum of squares by the degrees of freedom and therefore may be viewed as variance estimates. From Eq 25, the F statistic is the ratio of the between-groups estimate to the within- groups estimate, which, in this case, gives a calculated F of 4.1048, which, being greater than 3.17, falls within the critical region. In other words, results support rejection of the null hypothesis and the conclusion that bond strengths of the three coatings are not equivalent. If F had fallen outside the critical region, it would have been untenable to infer any difference in bond strength among the three coatings. Table 4 One-way analysis of variance of bond strength for three coatings 2 Mean ( ) N Variance (S ) 20 3000.89 Coating 1 204.5 20 2124.43 Coating 2 228.7 20 4162.43 Coating 3 178.3 203.5 60 3421.76 Total Partition of variance sum of squares Sum of squares Degrees of freedom Mean square F 25,416.3 2 12,708.15 Between-groups (BSS) 176,467.3 57 3095.92 4.1048 Within-groups (WSS) 201,883.6 59 Total (TSS) This discussion and example of analysis of variance deals with the simplest possible case. Analysis of variance can incorporate more than one categorical independent variable, and there exists a wide variety of specialized applications of analysis of variance to specific, complex experimental designs. Reference 7 provides a clear and accessible discussion of the extension of analysis of variance to two independent variables. More definitive and complete sources on analysis of variance are provided by Ref 9 and 10.

References cited in this section 7. H.M. Blalock, Social Statistics, revised 2nd ed., McGraw-Hill Book Company, 1979, p 205–206, 208– 218, 231, 299–325, 340, 348, 352–369, 388, 398, 554–551 9. H. Scheffe, The Analysis of Variance, John Wiley and Sons, Inc., 1959 10. J. Neter, M. Kutner, C. Nachtsheim, and W. Wasserman, Applied Linear Statistical Models, 4th ed., WCB/McGraw Hill, Inc., 1996, p 296–315, 531–566, 663–1041


Correlation and Regression Pearson's r. Previously in this article, consideration was given to statistical hypothesis testing. There is another class of statistical measures called measures of association. Correlations or measures of association measure dependence between two variables, or the ability to predict one variable from another. There are, in fact, many such measures designed for nominal and ordinal scales, and many of these are described in Ref 3 and 7. However, one particular measure of association, the Pearson product-moment coefficient, or Pearson's r (r represents the estimate; the symbol for the parameter is ρ), is probably the most widely recognized and used measure for two-interval or ratio scales. To begin a discussion of correlation, consider the expression for covariance in Eq 26: (Eq 26) Note the similarity to Eq 3 and 4 for the variance, and that, while the variance is necessarily a positive quantity, the covariance can be either positive or negative. Also note that the covariance reaches a maximum (positive or negative) when X and Y are perfectly correlated, and that it approaches zero to the extent that there is no association between X and Y. However, covariance is of more theoretical than practical interest as a measure of association. However, a number of other important and useful measures are based on covariance. One of these is Pearson's r:

(Eq 27)

In other words, Pearson's r is the ratio of the covariance of X and Y to the product of the standard deviations of X and Y (Ref 7). Unlike the covariance, r has a range of -1 to +1. Linear Regression Analysis. Figure 2 presents a scatter diagram. The variables in the scatter diagram are reflectivity measurements taken on reflective highway markers when new (X) and after weathering in service for one year (Y). As should be evident from the pattern, there is a fairly strong positive correlation of r = 0.900. In other words, the effects of weathering on reflectivity appear to be somewhat predictable. One expedient way to make that prediction is with linear regression analysis. Linear regression is based on the linear equation: (Eq 28) where β is the slope of the line, α represents the Y-intercept, or the value of Y when X = 0, and εi represents an error or disturbance term. The regression line is fit to the data according to a specific criterion known as least squares. The regression line that minimizes the squared deviations of the observed values of Y from the values predicted for Y by the regression line satisfies the criterion of least squares (Ref 11). As it turns out, this criterion is met if a and b are obtained as in Eq 29 and 30:

(Eq 29)

(Eq 30) where a and b are estimates for the parameters α and β. Regression analysis can be extended to cases involving more than one independent variable. There are also a variety of ways to model nonlinear associations using regression analysis. However, these topics are beyond the scope of this article, and the reader is referred to Ref 10 and 11 for more complete treatments of these and a variety of other extensions of simple regression analysis.

Fig. 2 Reflectivity of highway markers when new and after one year in service. Units for reflectivity are specific intensity (millicandela) per unit area. Regression Analysis and Statistical Inference. Regression analysis can be used to predict one variable in terms of other variables, but the terms in the regression equation itself tell the researcher very little about the dependability of the prediction. Statistical tests are available to test for significance of the entire regression equation and also for the significance of b. It is also possible to place a confidence interval around the predictions of Y obtained from a regression analysis. Each type of statistical inference is discussed in turn. However, first, there are a number of assumptions required for regression analysis that deserve consideration. To begin, there are specific assumptions relating to the disturbance term, ε. This term, also called the residual, represents measurement error in Y and specification error, or influences on Y not included in the regression equation. Regression assumes that ε has an expected value (mean) of 0, is approximately normally distributed,

and that the independent (X) variable is measured without error. Note that the estimated residual always has a mean of 0, but the assumption refers to the parameter ε. It is also necessary to assume that X and ε are statistically independent. Suppose that an unidentified third variable, Q, is the true source of influence on Y, but Q is also correlated with X. In that situation, X only appears to influence Y because of the influence of Q on both X and Y. This would be a case of violation of the assumed independence of X and ε. It should be noted, however, that an analogous assumption applies to all the statistical tests considered so far. Regression analysis also assumes a multivariate normal distribution. In the case of two variables, a bivariate normal distribution could be represented graphically as an elliptical bell in three dimensions, to the extent that a normal distribution is represented by a bell curve. In other words, for each value of Y, the Xs are distributed normally (Ref 7). Related to the assumption of a multivariate normal distribution is the assumption of homoscedasticity, which refers to the requirement that Y be not just normally distributed about each value of X but also that the variance of Y be consistent across values of X. Given these assumptions, discussion can turn to statistical inference in the context of regression analysis. For the sake of brevity, this discussion is also limited to the case of regression with one independent variable. In the case of simple regression with one independent variable, the test for the significance of β is identical to the test for the overall significance of the regression equation. The test parallels that for analysis of variance. Consider the following expression: (Eq 31) where Ŷi is the value predicted for Yi from the regression equation. Note that the term to the left of the equals sign is the total sum of squares. The first term to the right represents the residual, or error, sum of squares and is analogous to within sum of squares in the context of analysis of variance. The final term is the sum of squares that is accounted for or “explained” by the regression equation and is directly analogous to between sum of squares. This term is often referred to as explained, regression, or model sum of squares. At this point, it should be evident that an F ratio can be obtained from these quantities, given the appropriate degrees of freedom. Also analogous to analysis of variance, degrees of freedom for the regression sum of squares is (k - 1) and for residual sum of squares (n - k), except that in this case, k represents the number of terms estimated for the regression equation, in this case, a and b, or 2. Table 5 works through a numeric example based on the reflectivity data in Fig. 2. Note that an F ratio of 246.6 is obtained, and the critical value for F, assuming 1 and 58 degrees of freedom and α = 0.05, is approximately 4.01, so the null hypothesis that the coefficient of determination Ρ = 0 can be rejected. Table 5 Regression analysis summary for the reflectivity problem in fig. 2 Number of cases (N) 60 229.33 Mean reflectivity when new (X) Mean reflectivity after one year in service (Y) 206.52 -0.261 Constant (a) 0.9016 Slope (b) 0.900 Pearson's r 2 0.810 Pearson's r Source Sum of squares Degrees of freedom Mean square F 1 154,328.31 246.61 Regression 154,328.31 36,296.67 58 625.81 Residual 190,624.98 59 Total It is important to note that the foregoing test applies to the regression equation for both simple and multiple regression. However, it is applicable to the significance of β only in the limited case of regression with one independent variable. Another similar test is available for the significance of the individual estimates for the βs in multiple regression, but that is beyond the scope of this article. Coefficient of Determination. In the case of multiple regression, there is a measure called the multiple correlation coefficient, or coefficient of determination, known as R2. Note the use of the upper case “R” (upper

case rho, or Ρ, to denote a parameter) to differentiate it from Pearson's r. R2 is analogous and, in the twovariable case, equivalent to Pearson's r2. Consider Eq 32:

(Eq 32)

Referring back to Eq 31, R2 is also the ratio regression/total sum of squares. The coefficient of determination is a useful summary measure, because it expresses the results of the regression equation in terms of proportionate reduction in error, or the proportion of variance in Y that is accounted for by the regression.

References cited in this section 3. S. Siegel, Nonparametric Statistics for the Behavioral Sciences, McGraw-Hill Book Company, 1956, p 25–26, 30–31, 195–239 7. H.M. Blalock, Social Statistics, revised 2nd ed., McGraw-Hill Book Company, 1979, p 205–206, 208– 218, 231, 299–325, 340, 348, 352–369, 388, 398, 554–551 10. J. Neter, M. Kutner, C. Nachtsheim, and W. Wasserman, Applied Linear Statistical Models, 4th ed., WCB/McGraw Hill, Inc., 1996, p 296–315, 531–566, 663–1041 11. N.R. Draper and H. Smith, Applied Regression Analysis, John Wiley and Sons, Inc., 1966, p 58, 104– 158, 263–298


References 1. S.S. Stevens, On the Theory of Scales of Measurement, Science, Vol 103, 1946, p 677–680 2. W.D. Callister, Materials Science and Engineering: An Introduction, 3rd ed., John Wiley and Sons, Inc., 1994, p 132 3. S. Siegel, Nonparametric Statistics for the Behavioral Sciences, McGraw-Hill Book Company, 1956, p 25–26, 30–31, 195–239 4. W. Mendenhall and J.E. Reinmuth, Statistics for Management and Economics, Duxbury Press, 1974, p 192 5. N.T. Kottegoda and R. Rosso, Statistics, Probability, and Reliability for Civil and Environmental Engineers, McGraw-Hill Book Company, 1997, p 15–16, 222, 245–257

6. D. Salsburg, The Lady Tasting Tea: How Statistics Revolutionized Science in the Twentieth Century, W.H. Freeman and Company, 2001, p 25–32 7. H.M. Blalock, Social Statistics, revised 2nd ed., McGraw-Hill Book Company, 1979, p 205–206, 208– 218, 231, 299–325, 340, 348, 352–369, 388, 398, 554–551 8. W.G. Cochran, Sampling Techniques, 3rd ed., John Wiley and Sons, 1977, p 24–25 9. H. Scheffe, The Analysis of Variance, John Wiley and Sons, Inc., 1959 10. J. Neter, M. Kutner, C. Nachtsheim, and W. Wasserman, Applied Linear Statistical Models, 4th ed., WCB/McGraw Hill, Inc., 1996, p 296–315, 531–566, 663–1041 11. N.R. Draper and H. Smith, Applied Regression Analysis, John Wiley and Sons, Inc., 1966, p 58, 104– 158, 263–298

Copyright © 2005 ASM International®. All Rights Reserved. < Previous section in this article < Previous section in this article

Next section in this article >



“Applying Statistics to Analysis of Corrosion Data,” G 16, ASTM R.C. Rice, Statistics and Data Analysis, Mechanical Testing, Vol 8, Metals Handbook, 9th ed., American Society for Metals, 1985, p 621–720 “Statistics—Vocabulary and Symbols—Probability and General Statistical Terms,” ANSI/ ISO/ASQC 3534-1, The American Society for Quality Control

W. Reitz and J. Rawers, Materials Basics for the Corrosionist, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 980–991

Materials Basics for the Corrosionist Wayne Reitz, North Dakota State University; James Rawers, Albany Research Center

Introduction THIS ARTICLE is an overview of the science and engineering of materials for corrosionists with limited knowledge in this field. The presentation is descriptive, not technical, and assumes a basic understanding and knowledge of science. To help with the understanding of materials, definitions and descriptions of many of the more commonly used materials terms are presented along with examples. Advanced information can be found in the Selected References. Materials terms are printed in italics. If not immediately provided, the definitions can be found in the glossary at the back of this Volume. The first section in this article, “Classification of Materials,” lists and defines the categories of materials and gives examples of commercial materials. In general, metals are versatile materials that can be modified easily to achieve specific properties. Polymers can be produced readily in complex near-net final shape, and they have high ductility and low strength. Ceramics tend to have high compressive strength and good resistance to oxidation and corrosion; ceramics also exhibit low electrical and thermal conductivity and are brittle. The second section “Properties,” discusses the more frequently cited, measured, and relevant physical properties of materials, such as phases, strength, conductivity, and wear. The more common methods used to determine these properties and the standard set of units used are presented. The third section of this article, “Processing,” describes methods used to fabricate materials into different products. It includes a summary table of materials and their properties. The fourth and final section, “Material Design, Application, and Failure,” addresses material design, materials applications, and materials failure analysis. A collection of representative materials and their properties and characteristics is tabulated.

W. Reitz and J. Rawers, Materials Basics for the Corrosionist, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 980–991 Materials Basics for the Corrosionist Wayne Reitz, North Dakota State University; James Rawers, Albany Research Center

Classification of Materials Materials are defined as the elements, constituents, or substances of which something is composed or made. The study of materials can be divided into three overlapping fields: • • •

Materials science, concerned with the basic understanding and characterization of materials Materials technology, concerned with the development and processing of materials Materials engineering, concerned with the design and application of materials.

These fields have more commonalities than differences. While the materials scientist may be more concerned with developing fundamental theories, the technologist with applying those ideas to developing or improving

materials, and the engineer with taking these materials and building something from them, all three use common terminology. Materials are either solids or fluids. Fluids (gases and liquids) are the principal agents of corrosion. The characteristics and properties of fluids, with the exception of glasses, are beyond the scope of this article and are not discussed. Solids have relatively fixed volumes and are able to support shear loads. Engineered solids are the components of structures. The properties of the solid determine the design and construction of products. Processed solids can be divided into at least three categories: ceramics, metals, or organics. These categories are not exclusive and are roughly related to the type of bonding that holds the atoms together and their fundamental structural units as shown in Fig. 1 and described in Table 1. Ceramics have a large component of ionic bonds holding atoms together. The basic structural units of ceramics are molecules, which are rigid and can vary from 2 to 10 atoms in size. Organics have a large component of localized covalent bonds between atoms, and their basic structural units, again molecules, can vary from approximately 100 to more than 10,000 atoms. Organics are very flexible. Metals have a mutual sharing of electron bonding between all atoms. Their basic building unit is the single atom. In ionic bonding, one atom gives up its electrons to another atom; in covalent bonding, two atoms mutually share electrons. In metallic bonding, a large array of atoms give up individual ownership of their valence electrons, sharing them mutually.

Fig. 1 Fundamental structural units of materials. (a) Metallic bond showing a cluster of cations (+) in a sea of electrons (-). (b) Geometric isomers of polyisoprene, typical of organics Table 1 Material bonding Characteristics of solid materials Type of solid material Electronic bond Fundamental unit Shared between all atoms Single atom Metals Ionic bonding between atoms 2 to 10 atoms in rigid structure Ceramics Covalent bond between two atoms 100 to over 10,000 atoms Organics Varies by constituents Matrix and reinforcement Composites Many properties of solids are ultimately related to the interaction of atoms. Atomic interaction is explained by (a) electron interaction, which is best explained using quantum mechanics and (b) structural interaction, which

predominates as the number of atoms increases and can best be explained using Newtonian or continuum mechanics. The three most widely used engineering materials are concrete, wood, and steel. These materials are from each of the materials categories: ceramics, organics, and metals, respectively. They are widely available, economical to produce, can be processed in a number of different ways to produce a great variety of products, and thus find a large number of applications. Each material has its own set of properties that make it best suited for particular applications. For example: Application Requirements Example of material that meets most or all requirements Large quantities of material able to withstand large Concrete Dams compressive forces Low cost, on-site processing, and renewable Wood Home materials construction High strength, flexibility, and corrosion resistance Steel Large cargo to a salt water environment ships All materials have assets and limitations, such as availability, cost, strength, and corrosion resistance, which must be evaluated. Engineering guidelines and industry standards, as well as legal requirements, must be met when choosing a material for particular application. The materials information contained herein represents typical information and data. Consult the Selected References for more specific details or property values. Ceramics are composed of carbide, oxide, nitride, sulfide, and/or silicate molecules. Examples include: SiC and Si3N4, used for cutting tools; SiO2, the principal component in window glass, 2CaO·SiO2, one of the principal components in cement; and Al 2Si2O3 (OH)4, clay used in making bricks. Ceramics are characterized by having good compressive strength and low tensile strength. The ionic nature of the interatomic bonds produces high rigidity (stiffness) and little ductility (elasticity). Ceramics are strong; once a crack starts, it grows rapidly, leading to immediate and rapid failure. The ionic bond is a high-energy bond, more stable than either a metallic or covalent bond, thus forming relatively stable molecules. Ceramics are generally more corrosion and oxidation resistant than either metals or organics. The nature of the ionic bond also limits the mobility of the electrons in the molecule, and ceramics thus have limited electrical and thermal conductivity and make good insulators. Organics are composed of carbon-base materials. When carbon bonds covalently to other atoms such as other carbon atoms, nitrogen, oxygen, sulfur, and/or phosphorus, it can form extended one-dimensional chains (such as DNA molecules), two-dimensional (2D) arrays (graphite), or three-dimensional (3D) structures (diamond). Carbon atoms can share electrons or bond covalently up to four of its electrons with other atoms. As the degree of interlinking between carbon atoms increases, the structure becomes more rigid and the strength becomes greater, as shown in Fig. 2.

Fig. 2 (a) Simulated cross linked (networked or three-dimensional) structure. (b) Simulated linear polymer molecule. (c) Simulated structure of a branched polymer Polymers are repeating chains of organic units (known as mers or monomers) that are covalently linked. Plastics are processed polymers and are categorized by how they are formed. Thermoplastics have very long chains that, when heated, readily flow past one another. However, as the temperature is reduced, the thermal energy of the molecule is reduced, and the interatomic motion and bonds become more restricted. The shape of the polymers thus becomes more rigid. As polymer length increases, the polymers can become ever more intertangled, resulting in a material of limited strength but great flexibility. Some plastics are able to stretch several hundred percent before failure. Reheating can reshape these plastics. Thermoplastics include natural rubber, nylon, polyvinyl chloride, polyesters, wood, silk, and cotton. Thermosets are polymers with multiple covalent links between polymer chains. The greater the number of links between the different chains, the more restrictive the movement of individual polymers, resulting in greater rigidity and higher strength. Heating cannot reshape these plastics once they are formed. Thermoset plastics include phenolics, polyesters, epoxies, urethanes, Bakelite, polyvinyl chloride (PVC), and polyethylene. The localized nature of covalent bonds restricts the overall electron mobility in the polymer structure, resulting in high electrical resistivity and limited thermal conductivity. Organics generally have good corrosion resistance to mild acids and bases. However, solvents such as acetone attack organics, alcohol can dissolve polymer and interpolymer bonds, and ultraviolet radiation such as sunlight can break polymer bonds. Metals are composed of those elements that readily share their electrons globally with other atoms. The elements that have this ability constitute the majority of the periodic table. The exceptions are the noble elements, the halides, carbon, nitrogen, oxygen, phosphorus, and sulfur. The metallic elements are able to form metallic compounds or alloys by combining with different metallic elements. Metal atoms combine to form definite structural atomic arrangements or lattices, 3D crystal structures. Most metals have either face-centered cubic (fcc), hexagonal close packed (hcp), or body-centered cubic (bcc) unit

cells, as shown in Fig. 3. Metals with the fcc lattice structure have the highest ductility and lowest strength. The hcp metals have the highest strength and lowest ductility, and bcc metals are generally in between. When different metallic elements are mixed they can form a solid solution (random dispersion of atoms within a single phase), or they can: (a) form a different phase, (b) begin to segregate or cluster forming a second phase with separate or distinct composition and/or lattice, or (c) develop a very limited stoichiometric chemical relationship and form definite structure coordinated atomic configurations (producing intermetallics that have narrow chemical compositions). Because metallic electrons are able to move freely throughout the alloy, metals have good electrical and thermal conductivity.

Fig. 3 Unit cells and atom positions for metal lattices. The positions of the atoms are shown as dots at the left of each pair of drawings, while the atoms themselves are shown close to their true effective size by spheres or portions of spheres at the right of each pair. (a) Face-centered cubic (fcc). (b) Hexagonal closepacked (hcp). (c) Body-centered cubic unit cells (bcc) Semiconductors are materials that when pure are electrical insulators. For example, pure, single crystal silicon has a lattice structure similar to that of carbon diamond, is largely covalently bonded, and is an insulator. However, when minute quantities of impurities are present in the material (doped), the local electronic structure can be increased or depleted. This electron unbalance can greatly reduce the resistivity of the material, thus creating a semiconducting material. Composites are combinations or mixtures of different phases or discrete materials. The range of composite possibilities is almost unlimited in composition and scale, as listed in Table 2. Composites are desirable because they are structures with properties of the individual components. For example, it is possible to increase the

strength of metals while retaining much of the ductility of metal by incorporating and combining into the metal matrix very fine ceramic particles. Steel is a mixture of bcc iron with a dispersion of iron-carbide ceramic particles. Polymers can be strengthened with the addition of metallic wires (steel-belted automobile tires). Ceramics can be given some ductility by placing them in a metal matrix, for example, tungsten carbide coated cobalt cutting tools. Plywood is a composite in which the wood laminates are bonded by epoxy polymers. Strength properties of wood are highly directional. Laying-up the different layers of the plywood in different directions improves the overall directional strength of the final product. Care must be taken in designing and using composites. Just like the plywood, reinforced materials tend to be anisotropic their properties varying with direction. The choice of materials that results in improved composite properties in one application may possess undesirable properties in another application. For example, while the strength of a composite may be enhanced, the corrosion resistance may be decreased. Table 2 Composite types Type Metal-metal

Metal-organic

Metal-ceramic

Organic-organic

Organic-ceramic

Example, common name Components and purpose Galvanized steel Fe: bulk properties such as formability and strength

U.S. coin (penny)

Zn: surface coating for corrosion protection Zn: structural material

Painted steel structure

Cu: surface coating (modification) Fe: strength

Electrical wires

Paint: corrosion protection Cu: electrical conductivity

Cermet

Organic: electrical insulator Co: ductile binder

Reinforced concrete

WC: hard cutting component Cement: binder, compressive strength

Particle board

Fe-rods: ductile, tensile strength Wood: strength

Fiberglass

Epoxy: binder Glass: strength

Paper

Resin: binder Cellulose: structural material

Ceramic-ceramic Concrete

Porcelain

Clay: surface finish (modification) Gravel: compressive strength Cement: binder Ceramic: structural material Glaze: surface finish (modification)


Properties The ASM Handbook series is a good source of information about materials properties. The primary source of information for evaluating materials properties or materials testing is ASTM International (formerly the American Society for Testing of Materials), an organization devoted to establishing standardized procedures for all types of testing. The International Organization for Standards (ISO) promotes standardization and related activities worldwide. The work of the ISO results in international agreements that are published as International Standards. There are also regional standards such as the EN (Euronom) and other national standards organizations that are used to specify materials, materials testing, and quality. Composition and phases are the most important materials properties. In principle, all other properties can be derived from knowledge of these two. The electronic interaction or bonding between the atoms can be determined from knowledge of the chemical elemental composition (the interatomic positions). Composition can be obtained from analytical chemistry (bulk chemical analysis), emission spectroscopy, x-ray fluorescence, and Auger spectroscopy. The measuring units for chemical composition are atom or mole fraction for scientists and weight percent for engineers. The structure of a material is its atomic arrangement and lattice spacing. From knowledge of the atomic placement and the increased complexity that arises as the number of atoms increases, there emerge new theories controlling material behavior. Structure can be determined using x-ray diffraction. Other methods include ultrahigh resolution electron microscopy and atomic microscopy. These methods are used to determine local (molecular structure), long- range order (lattice structure), the presence of local defect structure, or total disorder (amorphous structure). Most ceramic and metallic materials have an ordered structure whereas most organic materials and glasses have limited or no atomic order. The size of the fundamental building component of a solid influences the resulting microstructure greatly, as shown in Fig. 4. The basic unit for most metals is the individual atom. Because the bonding is due to a global sharing of electrons, the atoms can be considered to have spherical symmetry and thus no orientation preference. The packing structure of metals is similar to the packing of tennis balls, in that atoms are stacked so that each is part of a repeating 3-D pattern. The unit cell is a coordinated set of atoms that form the 3D lattice that repeats by a simple translation. The basic unit for ceramics is the individual molecule, for example, SiO 2 (Fig. 4a). Molecules are a set of bonded atoms that have a rigid structure and a more complex geometry than single atoms. The bonding between the molecules is not as strong as the bonding within the molecule. The additional complexity associated with the molecular shape makes the unit structure more difficult to define than that for simple atoms. Many ceramics align readily into a lattice structure similar to that of metals with the individual metal atoms replaced with the ceramic molecules. However, rapid solidification or the presence of impurity molecules of different shape can result in a solid structure with limited local alignment of molecular unit cells, which, in the extreme, can result in amorphous or glass structures. The basic unit for organics is the polymer. Polymers have large complex atomic alignments, irregular geometry, and flexible molecules (Fig. 4b). Most organic solids have an amorphous or glass structure, as defined in Table 3.

Fig. 4 The unit cells of a ceramic and a polymer. (a) Idealized cristobalite structure (b) Stereoisomers of polypropylene; from top: isotactic, syndiotactic, and atactic

Table 3 Characteristics descriptors of solids Term Description Nonrepeating arrangement or alignment of fundamental building units Amorphous 1, 2, or 3D alignment or repeating arrangement of fundamental building units Crystalline Region of space with uniform composition and uniform structure throughout Phase Single-grain, uniform composition, same structural alignment throughout Single phase Structure composed of a number of single phase crystals Multiple grain Grain boundary Interface or boundary between two grains Multigrain structures composed of different phases Multiple phase Second-phase smaller grains within large grains Precipitates Microstructure and surface morphology can be determined by optical and electron microscopy. Optical microscopy is typically used for 10 to 1000 times magnification. Numerous sample preparation and microscopic techniques have been developed for examination of flat polished surfaces to determine grain structure and, possibly, identify phases. Light or optical microscopy permits color to be observed and is used often to identify and differentiate microscopic features. Low magnification is used often for failure analysis to show the type of failure: intergranular or transgranular. With optical microscopy, the higher the magnification, the lower is the depth of field resolution, (focus in the vertical direction). Electron microscopy, especially scanning electron microscopy (SEM), has essentially an infinite depth of field and can magnify in excess of 100,000 times. Scanning electron microscopy is ideally suited for examination of surface features. Many SEMs are equipped with auxiliary features that allow for phase differentiation (secondary electron detection), atomic number differentiation (backscattered electron detection), quantitative and qualitative chemical analysis using energy dispersive spectroscopy (EDS), wavelength dispersive spectroscopy (WDS), and even rudimentary xray diffraction analysis (backscattered Kikuchi line analysis). The phase of a material is defined as a region with homogenous atomic structure and chemistry; for example, amorphous silicate glass or a long-range repeating lattice of bcc iron. Many materials are composed of multiple phases (regions with different chemical composition and/ or different structures). The energy differences in phases may result in localized galvanic corrosion. Rarely does an entire piece of material maintain homogeneous structure. Grains (or crystals) are regions within a material that have both homogenous chemistry and atomic structure. Precipitates are small grains of different chemistry within a large grain. If the changes in homogeneity or local structure of the material are abrupt, the discontinuity results in the formation of a grain boundary. Grain boundaries may result from a change in the direction of atomic orientation, from difference in chemical composition, or from different phases being present. The abrupt property changes at a grain boundary are regions of energy higher than the energy level within the grain. This results in a more corrosion-prone region, a region along which diffusion is more rapid, and a region of higher stress. In materials with a lattice structure, such as metals and ceramics, a line dislocation is another type of lattice structure that can be present. Dislocations are missing lines of atoms within grains and are important in understanding the plastic or ductile nature of a material and the increase in a material strength as it undergoes plastic deformation. Line dislocations are areas of higher localized stress. Micrographs of surface morphology and microstructures are shown in Fig. 5.

Fig. 5 Microstructure of materials. (a) High-strength low alloy steel (0.2% C) hot rolled. The structure is ferrite and pearlite. 4% picral, then 2% nital etchants were used. Magnification is approximately 200×. (b) 1045 steel sheet, 3 mm (0.13 inch) thick, normalized by austenitizing at 1095 °C (2000 °F) and cooling in air. Structure consists of pearlite (dark gray) and ferrite (light). Picral. 500× (c) Alloy 242-T571, permanent mold cast and artificially aged. Structure contains blades of NiAl3 (dark gray) in the mediumgray Cu3-NiAl6 script. CuAl2 particles (light) and scriptlike Mg2Si (black) also are present. 0.5% HF. 250× (d) Graphite-silver copper composite (Thornel 300 fiber in 70Ag-30Cu eutectic matrix), unidirectional. Liquid-metal infiltration of fiber bundles followed by diffusion bonding. SEM of a failed tensile surface. The matrix shows good ductility, but the lack of matrix adhering of the fibers indicates low interfacial bond strength. 1860× As particle or grain size decreases, the structure eventually reaches a size where the number of atoms associated with surface or grain boundary of a particle begins to approach the number of atoms within the particle or grain. Such materials are referred to as nanomaterials and are defined as materials having one or more structural dimensions between 1 and 100 nm (10-7 to 10-9 meters). To date, understanding nanomaterials properties has

required no new physics, and most properties of nanomaterials can be explained using either quantum mechanics or classical mechanics. Atoms or molecules at a surface have different environments surrounding them than do atoms and molecules in the interior of the particle or grain. Thus, surface effects— surface physics, chemistry, and mechanics, are important in understanding many of the phenomena associated with nanomaterials. The relationships between composition, temperature, and phase often are presented in phase diagrams. These graphs describe the phases present as a function of the composition (most often wt%) on the abscissa and temperature as the ordinate. Equations have been derived from which the relative concentration of the different phases present as well as the composition of each phase present can be derived, given a temperature and an overall material chemical composition. Phase diagrams are derived for systems at thermodynamic equilibrium. Phase diagrams are generally limited to only two or three components because of the increase in complexity with increasing number of components, that is, atomic elements or molecules. Phase diagrams are extremely important in understanding the microstructure of a material and its properties. The iron–iron-carbide phase diagram (Fig. 6) describes carbon steels and is well known in the metals community.

Fig. 6 Iron-carbon phase diagram

Mechanical Properties. One of the more important and most often measured sets of materials properties is response to an applied load; the strength (ability of a material to support a load) and ductility (deformation under a load). A normalized response characteristic of the bulk material strength property defined as stress and symbolized as s or σ, the applied load divided by the cross section of a material, is measured in Pascals (Pa = newton/meter2) or psi (pounds/ inch2). The conversion is that 1 psi is approximately equal to 7000 Pa. A normalized material change in shape defined as strain, symbolized as e or ε, unitless, but with units sometimes given as length/length, is the change in length of a material divided by its original length, δl/l, or change in area over area, δarea/area. Engineers measure the stress-strain curve with respect to the original unstressed length or area, whereas scientists use the instantaneous length or area. During the initial phase of stress loading the material is elastic; that is, if the stress load is removed, the material returns to its original shape. During this elastic response, the stress-strain relationship is linear. The stiffness of a material, rigidity, or modulus of elasticity, E, is obtained by measuring the slope of the elastic stress-strain curve. Elastic modulus is measured using the same units as stress. Fig. 7(a) shows a typical stress-strain curve, and Fig. 7(b) shows typical curves for materials with different behaviors.

Fig. 7 Mechanical properties diagrams. (a) Typical stress-strain tensile curve (b) Typical stress-strain tensile curves for materials with different characteristics. As the stress load increases, the material reaches a point at which it either breaks or fails, or yields and becomes plastic. Plastic deformation or yielding is defined as a permanent change in shape of a material in response to an applied load. After yielding, the change in strain per unit load increases. The stress at which the material becomes plastic is defined as the yield stress, σys. Engineers normally design for loads less than σys. Ceramics generally have very limited ductility, and the yield stress is often the failure stress. However, during fabrication, materials are often required to be deformed to a prescribed shape and thus must be loaded above the yield point. As the material undergoes plastic deformation, it becomes stronger due to (a) the generation and entanglement of dislocations within each grain in metals or (b) the alignment of polymer chains in organics. If the material is unloaded prior to failure, the material responds elastically from the point of maximum load. On reloading, the plastically deformed material will have a higher yield stress and is said to have undergone work hardening. The material can be returned to its original yield stress by heating (recrystallization). Deforming the material at elevated temperatures reduces the stress needed for deformation and eliminates the work hardening (hot working). In metals, regions of high tensile stress are anodic compared with regions of lower stress. A bent piece of tubing possesses higher residual tensile stresses on the outside of the bend. Material response to load depends on the type of material, temperature, type of loading, previous loading, and the rate at which the load is applied. The most often cited stress-strain data are derived from tensile tests. In this test the response of a material to increasing pull or tensile stress is evaluated (normally at a strain rate of 0.02/s) at room temperature on samples with a typical sample-test length of approximately 5 cm (2 in.) and cross section diameter of approximately 1.25 cm (0.5 in.). Another commonly used stress-strain test is the

compression test in which a cylinder is compacted. While metals have similar response to tensile and compression loads, ceramics generally have significantly higher compression strength than tensile strength and the strength values often are only reported for compression. Shear tests are also significant. Strength of metals is strongly influenced by grain size (size of crystals). Strength increases as grain size decreases. The decrease in grain size is accompanied by an increased volume fraction of grain boundaries. The grain boundaries are high-energy regions that are capable of accelerated corrosion, which is the basis for etching for metallographic examination. Grain boundaries act as sinks for impurities, which can also accelerate corrosion. Elevated temperature processing encourages grain growth, thus reducing the volume percent of grain boundaries. Another common measure of material response to a load is the hardness test, where the resistance of a material to a specific load using an indenter is measured. A number of different loads and indenter sizes, shapes, and materials have been adopted as standards. The area or depth of the indent is measured and reported as the Brinell, Rockwell, Knoop, or Vickers hardness number. In general, as the hardness of a material increases, so does its strength. An empirical relationship has been established between hardness and tensile strength. In practice, many materials contain flaws or defects. Material imperfections can have serious consequences on the response of a material to a load by reducing the stress to fracture and/or changing the nature of the failure. Standards have been derived to measure the change in failure if surface cracks are present. Fracture toughness (symbolized by KIc, KIIc, or KIIIc, depending on load configuration) provides a relationship between catastrophic failure and the size of the surface crack. Fracture toughness is valid only for materials that undergo brittle rapid failure with little or no plastic component prior to failure; that is, only elastic deformation followed by failure (defined as plane-strain failure). For materials with a significant plastic component prior to failure (defined as plane-stress failure), the area under the elastic and plastic portion of the stress-strain curve, ∫σδε, is the failure energy per unit volume. It is measured and defined as the fracture energy, (symbolized as JIc, JIIc, or JIIIc). In this type of failure, the advancing crack tip is often observed to have undergone plastic strain, and the sharpness of the crack tip is blunted and undergoes plastic deformation. Other material strength measurements include the response of a material to long-term loading, dynamic or impact loading, and changing loads, as defined in Table 4. Materials at constant stress levels below the yield strength over extended periods of time exhibit permanent deformation or creep and fail at loads below those measured in tensile tests. This is especially true for plastics. Materials loaded at high strain rates; that is under dynamic or impact loading, generally have higher yield strength but less ductility prior to failure than that measured in tensile tests. Materials subjected to constantly changing cyclic loads also undergo plastic deformation over time at loads much less than the yield strength and will fail in fatigue after a number of cycles. Fatigue curves are typically called S-N curves (stress versus number of cycles). Since the long- term properties of materials are important in typical service applications, numerous handbooks provide tables with creep and fatigue data. Figure 8 shows a typical creep curve and a fatigue curve. Table 4 Solid material mechanical properties Material strength Creep Compressive Tensile Shear

Impact Fatigue

Description Material response of deformation over time to a constant load. Creep-rupture strength is the stress that will cause fracture in a given time. The maximum compressive stress a material can develop. Compressive stress results in deformation (shortening) in the direction of the applied force. The maximum load a material can support in response to a pulling force at a slow strain rate. The ultimate strength as contrasted to the yield strength The maximum stress a solid can sustain when the force is tangential to the plane on which it acts. In metals the stress exists when parallel planes of metal crystals slide across each other. In plastics, layers of molecules slide across each other. A measure of resiliency or toughness of a solid to a shock or very rapid strain rate. The amount of energy that can be absorbed without fracture Material response leading to fracture to variation in stress (or strain) loading at less than the ultimate strength. Often cyclic loading

Fig. 8 A response of a material to long-term loading (a) Typical creep curve showing the three stages of creep (b) Typical S-N curves for constant amplitude and sinusoidal loading Thermal Properties. Thermal conductivity is the ability of a material to transmit heat. In most applications, thermal conductivity is the result of coordinated atomic or molecular movement (phonon wave) that is enhanced by long-range order. Metals possess higher thermal conductivity than ceramics, which are considered to be thermal insulators, while amorphous plastics are even better thermal insulators than ceramics. High thermal conductivity enhances the ability of a material to remove heat in applications such as metal-cooling fins on air-cooled engines for lawn mowers and motorcycles. Ceramics and plastics make good coffee cups because they do not readily transmit heat to the surrounding atmosphere or a person's hand. Materials undergo thermal expansion as they are heated. Thermal expansion is the response of a material to an increase in individual, random atomic vibration. Increasing the internal energy of a material will cause the atoms or molecules to increase their vibration amplitude, increasing the distance between atoms. For materials that are packed closely, such as metals in a long- range lattice structure, the expansion is proportional to the change in thermal energy. The force a thermally expanded metal can exert is equal to the force it takes to expand it by the same amount. Ceramics and plastics have a more open structure resulting in a lower rate of thermal expansion than metals. Thermal expansion can induce compressive and tensile stresses within a component or assembly of components.

When a material is subjected to rapid heating, the combined effects of thermal expansion and thermal conductivity may result in serious structural problems. Metals, because of their high thermal conductivity and good ductility, are not as affected by the thermal strain produced by the differential thermal expansion as ceramics. Ceramics have much lower thermal conductivity than metals, and thus experience a sharper thermal gradient that can produce greater thermal stress due to differences in thermal expansion over shorter distances. This results in surface spalling or cracking and compression (during rapid cooling) or tensile (during rapid heating) failure. Plastics also have low thermal conductivity, but their enhanced elasticity greatly reduces chances of failure due to thermal shock. However, their low thermal conductivity may cause localized heating and weakening during fatigue loading. Diffusion is atomic or molecular movement within a material due to chemical gradients within the material. Processing materials at elevated temperatures and then cooling them rapidly before they can reach thermal equilibrium may result in compositional segregation, retention of unstable phases, retained thermal stresses, and so forth. Restoration of the system to thermodynamic equilibrium is accomplished by annealing. Many materials are processed in a manner that ensures the system is not in equilibrium. Most carbon steels have economic value because of their retention of the nonequilibrium Fe-Fe3C microstructure. The room temperature Fe-C thermodynamic equilibrium state would be a gray cast iron—that is, pure iron in a bcc lattice and graphite microstructure. Surface Interactions. Interactions at the solid surface with the environment can involve solid- gas interactions (oxidation), solid-liquid interactions (corrosion), and solid-solid interactions (wear). These interactions can be detrimental to the overall performance of the materials, as defined in Table 5. Table 5 Characteristics of common material loss interactions on solid surfaces Agent Degradation Description Oxidation Atmospheric reaction with solid surface to form solid or gaseous product Gas Liquid reaction with solid surface to dissolve or form new surface phase Liquid Corrosion Cavitation Gas bubble implosion on solid surface in liquid media Erosion Solid or liquid impact on solid surface Solid Abrasion Solid scraping across solid surface Oxidation is the chemical attack of gaseous elements (such as oxygen, sulfur, nitrogen, and chlorine) with metal surfaces. Most often the result of the interaction is the formation of a surface film that reduces or eliminates further solid- gas reaction. The type, density, thickness, and microstructure of the surface film depends on temperature, concentrations in the gas atmosphere, and thermodynamics of the reactions. The integrity and tenacity of the surface film to the metal surface depends on the difference in thermal properties, such as thermal expansion and thermal conductivity, as well as the difference in crystal lattice. Spallation is the flaking or chipping off of the surface film. Solid-gas interactions are normally evaluated by measuring the increase in material weight or the decrease in metal thickness due to development of the surface film with time. Thermodynamics of film compositions often are reported on a graph of temperature versus Gibb's free energy, Ellingham diagrams. Aqueous corrosion is the chemical attack by liquids of metal surfaces. Note that prior to exposure to liquids, most metal surfaces are coated with a thin oxide surface due to prior interaction with the atmosphere. Corrosion reactions result in the dissolution of the material into the liquid. In some reactions, selective leaching or preferential removal of selected elements within the metal occurs. The corrosion reactions may also produce a protective film or precipitate on the metal surface and reduce the rate of metal-liquid interaction. The rate of metal removed from the surface depends on the diffusion rate of the dissolved metal from the surface, the kinetics of the reaction, and concentrations of the corroding element and temperature. Corrosion rates are determined by long-term immersion in solution where the corrosion rate is a measure of the material mass loss or by short-term evaluation using electrochemical tests where the corrosion rate (corrosion current) is measured as a function of material potential. Thermodynamics of corrosion reactions are plotted on Pourbaix diagrams that show the dominant state of the metal ion as a function of pH and potential (see the article “Potential versus pH (Pourbaix) Diagrams” in this Volume). Wear is the physical removal of the surface of a material. Erosive wear is caused by solid or liquid particles impacting on the material surface. Cavitation erosion is caused by implosion of air bubbles formed on the surface of a solid as a result of rapid movement through a liquid (i.e., on a propeller of a ship). Abrasive wear is

caused by solid particles scraping across or impacting the material surface. Wear is extremely common and accounts for one-third to one-half of all equipment replacement in industry. Wear is a localized phenomenon, often the result of repeated events, and is characterized by chipping, plowing, or cutting away of the material at the surface. While individual event interactions can be almost innocuous, the cumulative effect over many events may remove a significant quantity of material and cause failure. Wear, like corrosion, is a complicated phenomenon and is influenced by a large number of factors. For example, erosion wear factors include hardness difference of the materials, relative speed, size, angle of impact, and morphology of the impacting particle and temperature. The combination of wear and corrosion often has synergistic effects on the failure of a material. At elevated temperatures, abrasive wear can prevent the buildup of the tough, wear resistant, surface oxide layer, thus providing a continuously fresh surface for oxidation. In a liquid media, corrosion or erosion may remove a tough, protective surface film, exposing the less corrosion and erosion resistant metal to continuous attack. Electrical, Magnetic, and Optical Properties. Electrical properties are related to the ease of moving electrons, ions, or charged molecules. This in turn is related to electron and atomic bonding and the material microstructure. Because metals share their valence electrons throughout the entire structure, there is little resistance to the ability of electrons to move. Conductivity is one of the distinguishing features of metals. Ceramics have ionic bonding, but the lattice structure does not allow for easy ionic movement. Thus ceramics are normally insulators. In organic compounds the covalent bonding between carbon atoms generally restricts the electron movement. An exception is graphite, which, because of the resonance between the interconnecting benzene rings, allows electrons to flow on a plane. Ionic and dipolar materials, while not providing good electrical conductivity, are used to store electrical energy by responding locally to electric fields that reorient their charge. This ability is utilized in capacitors and is a measure of the dielectric properties of the material. A change in shape or microstructure of a material due to an applied electric field is defined as the Piezoelectric effect. Superconductivity is the cooperative movement of a pair of electrons traveling through a lattice. This effect occurs readily at very low temperatures, from 0 to 70 K, when atomic vibrations are minimal. Magnetic properties are related to the interaction between the inner or core electrons of atoms. Elements with the strongest magnetic properties are those in the center of row four of the periodic table (with an unfilled dshell), row five (with an unfilled f-shell) and some of the rare earth elements. Magnetic molecules often are composed of iron-base spinel and sesquioxide compositions. Hard or permanent magnets are those materials that, once magnetized, readily retain their magnetic properties, for example, magnetic recording devices. Soft magnets are those materials that readily lose their magnetism, such as transformer cores. Material response to electromagnetic energy, transmission, reflection, or absorption is a function of the material electron bonding and atomic microstructure and the wavelength of the radiation. The electrons in metals are bonded very loosely and react readily with electromagnetic fields (e.g., radio antennas). The dipole structure of ceramics and many polymers will respond to certain electromagnetic frequencies. Note that in microwave ovens, the electromagnetic frequency is chosen to resonate with the vibrational energy of water molecules and not with carbon bonds of polymers or most ceramic dipoles. Thus, food is heated selectively and not the plate. If the irradiating wavelength is less than that of the grain structure, grain boundaries will reflect the light. Light transmission is also affected by abrupt changes within a material. Thus, the amorphous structure of many polymers and ceramics will produce transparent materials. However, small changes in microstructure will result in local changes in light transmission. The local alignment of polymer chains, glazing, in response to stress, reduces the amorphous local microstructure and produces a translucent region. Performance Maps. Material performance maps have been generated that relate material properties to environmental or engineering needs. (These are referred to often as Ashby maps). Generalized materials properties equations often can be separated into those variables that are intrinsic material properties and those variables that can be imposed upon the material by an external source. These equations present a theoretical or empirical model characterizing the response of a material. Often, as the external variable changes, a point is reached where the type or mode of material response changes; that is, the stress (external force)-strain (material response) changes with stress loading rates from creep to impact. Plots can be made using these equations with the material variable on one axis and the applied variables on the other axis. Boundaries can define regions or areas in which a range of material properties will respond in a similar manner to a range of imposed external conditions. Using these generic material performance maps, experimentally determined values for different materials can be derived. Often trends are observed between different types of materials or within a material

class that enables generalizations to be made where there may be no data to predict a material response to changing external forces. A typical Ashby map is shown in Fig. 9.

Fig. 9 Typical Ashby map. Strength, σf, plotted against density, ρ, for various engineered materials. Strength is yield strength for metals and polymers, compressive strength for ceramics, tear strength for elastomers, and tensile strength for composites. The guide lines of constant σf/ρ, /ρ, and /ρ are used in minimum weight, yield-limited, design. Note: log scales are used to encompass the range of values.


Processing Most materials do not come from nature ready for use. They must be refined, purified, alloyed, combined with other components, and shaped into a final product. The type of processing a material receives before application greatly affects the properties of the material and thus its applications. Representative mechanical and physical property values of materials from the different groups are found in Table 6.

Table 6 Summary of mechanical and physical properties of selected engineering materials Material name

Metals 1018 steel 1080 steel Cast iron Stainless steel Copper Zinc Brass Ti-6-4 2024 aluminum NiAl Ceramics Quartz Glass Portland cement Refractory brick Al2O3 TiC Organics Polyvinyl chloride Polyester Polyethylene Teflon Nylon

Symbol or composition

Yield strength MPa ksi

Ultimate strength MPa ksi

Elastic modulus GPa 106 psi

Elongation at failure

Fe-0.18C Fe-0.80C Fe-3.5C Fe-18Cr-15Ni

350 500 20 200

51 73 3 29

450 600 250 500

65.3 87.0 36.3 72.5

200 200 100 200

29 29 14.5 29

0.3 0.2 0.5 0.5

Cu Zn 70Cu-30Zn Ti-6Al-4V Al-4Cu

100 20 300 140 20

15 3 44 20 3

400 50 500 240 50

58.0 7.3 72.5 34.8 7.3

120 100 80 120 100

17.4 14.5 11.6 17.4 14.5

NiAl

250

36

1000

145.0

200

SiO2 SiO2-10CaO-13Na2O 63CaO-22SiO2-6Al2O33Fe2O3-3MgO (Al,Na,Mg)2(Si2O5)(OH)4

7200 1044 7200 36 5 36 30 4 30 4

1

Al2O3 TiC

20 250

-(-CH2-CHCl-)n… -(-CH2-CH2-)n-(-CFl2-CFl2-)n-(-CO-4CH2-CO-N-CH2N-)n-

Crystal structure

Electrical resistivity μΩ · Ω × m circularmil/ft

Thermal Thermal conductivity expansion W/m Btu/h 10- 10·K · ft · 6/K 6/°F °F

Toughness

Density

MPa

ksi

g/cm3 lb/in.3

bcc bcc bcc fcc

0.13 0.18 0.5 0.72

78 110 300 430

80 80 80 25

46.2 46.2 46.2 14.4

10 10 10 10

5.56 5.56 5.56 5.56

100 100 3 80

91 91 2.73 72.8

7.8 7.8 7.8 7.9

0.28 0.28 0.28 0.28

0.55 0.65 0.5 0.3 0.5

fcc hcp bcc hcp fcc

0.018 0.06 0.06 1.6 0.045

11 36 36 940 27

400 110 300 10 250

231.2 63.6 173.4 5.8 144.5

20 40 30 10 20

11.12 22.24 16.68 5.56 11.12

80 10 50 100 40

72.8 9.1 45.5 91 36.4

8.9 7.1 8.5 4.5 2.7

0.32 0.26 0.31 0.16 0.10

29

0.5

fcc/bcc

1.3

700

…

…

5

2.78

5

4.55

6

0.22

1044.0 74 5.2 100 4.4 100

10.73 14.5 14.5

0 0 0

Cryst Amor Amor

1017 1017 …

6 × 1019 6 × 1019 …

2 38 14.5

1.2 22.0 …

8 1 5

4.448 0.556 2.78

1 1 2

0.91 0.91 1.82

2.7 2.5 2.4

0.10 0.09 0.09

5

0.7

100

14.5

0

Amor

…

…

…

…

5

2.78

2

1.82

2.3

0.08

3 36

200 250

29.0 36.3

350 300

50.75 43.5

0 0

Cryst Cryst

1019 …

6 × 1021 …

20 26

11.6 15.0

6 7

3.336 3.892

4 4

3.64 3.64

3.8 4.5

0.14 0.16

45

7

60

8.7

4

0.58

1

Amor

1017

6 × 1019

0.2

0.1

70

38.92

5

4.55

1.4

0.05

40 15 20 50

6 2 3 7

80 30 50 70

11.6 4.4 7.3 10.2

4 1 0.5 3

0.58 0.145 0.0725 0.435

3 3 4 3

Amor Amor Amor Amor

1018 1012 1019 1019

6 × 1020 6 × 1014 6 × 1021 6 × 1021

0.2 0.3 0.2 0.2

0.1 0.2 0.1 0.1

70 70 100 80

38.92 38.92 55.6 44.48

5 5 5 5

4.55 4.55 4.55 4.55

1.4 0.9 2.2 1.2

0.05 0.03 0.08 0.04

-(C6H4-OH-CH2)nBakelite … Rubber … Wood, oak … Wood, pine Cryst, crystalline. Amor, amorphous

40 10 50 30

6 1 7 4

50 30 70 50

7.3 4.4 10.2 7.3

6 1 12 8

0.87 0.145 1.74 1.16

0.02 5 0.2 0.2

Amor Amor Amor Amor

… 1014 … …

… 6 × 1016 … …

0.1 0.1 0.4 0.4

0.1 0.1 0.2 0.2

20 150 … …

11.12 83.4 … …

0.1 5 1 1

0.091 4.55 0.91 0.91

1.5 1.2 0.7 0.3

0.05 0.04 0.03 0.01

Ceramics often are classified as being either low tech or high tech. Traditional or low-tech materials are those that are mined and, with little refinement, are ready for applications. Such materials include clays and silicates for bricks, pottery, concrete, and refractories. Advanced, or high-tech, ceramics are those with an application requiring a high degree of purity. Such materials include SiO2 for semiconductors and TiO2 for paints. Most ceramics processing begins with powders. Because of their high melting points, many ceramic products are sintered-bonded powders. In sintering, the ceramic powders are blended with binders or other low melting ceramic powders or formed into a slurry. The new composition is formed into a near final shape, or green product, and heated for sufficient time so that the particles either surface diffusion bond together or the lower melting ceramic melts and forms a thin layer that holds the higher melting particles together. In many ceramic products the final composition has a fine structure porosity as high as 6 to 10%. The ionic bonding of atoms in ceramics forms molecules that often have highly ordered localized structure. When these ceramics are cooled, they often do not go through a definite phase transition from liquid to solid. With decreasing temperature, the mobility of molecules becomes limited, and they may not be able to align themselves into a perfectly repeating lattice. The liquid viscosity continues to increase, eventually reaching a state where there is no molecular movement. This noncrystalline or amorphous structure is called glass. In commercial silica glasses, additives are placed in the SiO2 to reduce long-range order present in quartz, lower the melting temperature, and increase the viscosity. The resulting ceramic alloy can be formed readily into a final shape that will be retained at room temperature. By alloying, it is possible to reduce greatly the alloy melting point (Tm) beyond that of the individual elements or molecules. Commercial glass is made by alloying SiO2 with Tm ≈ 1700 °C (3100 °F) with Na2O having Tdissoc ≈ 1300 °C (2370 °F) and CaO, Tm ≈ 2600 °C (4710 °F). The resulting softening, or melting, temperature can be as low as 800 °C (1470 °F). Similarly, low melting glazes (e.g., CaO·Al2O3·2SiO2) are applied to the surface of ceramic green products (such as pottery and tableware) prior to firing so that, when heated, the glaze melts and forms a continuous, impermeable surface. The difference in thermal conductivity and thermal expansion between ceramics and metals makes bonding between these materials very difficult and often requires a series of intermediate steps. Because of the somewhat open structure of ceramics at low temperatures, organic glues can be used to bond ceramics to other materials. Because ceramics have good resistance to oxidation, many being composed of oxides and having low thermal conductivity, often they are used as coatings or protective barriers on metals for high temperature applications. In the ceramic- metal composite, the ceramic component offers protection from the environment while the metal component provides the necessary strength and resistance to catastrophic failure. Organics. The most common organic materials considered here are polymer-based plastics. These are long chain materials with carbon as their backbone. They are formed by reacting smaller size molecules called mers or monomers together to form polymers. Plastics are blends of polymers with different polymer lengths or with different polymer compositions. The structure of the polymer chains and length of the chains control many of the plastic properties. Monomers can be as simple as a two carbon molecule or can be very long carbon chains, which may contain small side chains. Polymerization is the reaction of individual similar monomers resulting in polymer of repeating similar units. In addition to reaction between similar mers, copolymerization occurs when the reaction is between different monomers. These reactions can be stimulated by temperature, by catalysts, or by reaction between the monomers themselves. Poly[chmerization and copolymerization are often incorporated into the final production step of plastic products. Individual polymers interact by directly linking between the different polymer chains, by weak van der Waals bonding between atoms on different chains, or by mechanical interlocking of the polymer chains. Plastics generally have an amorphous microstructure; however, chains may be aligned partially during fabrication by drawing the polymers through dies. Thermoset plastics are formed in final shape by placing the polymers or combination of polymers in a final shape die and then stimulating the reaction that initiates the cross linking between polymers. Thermoplastics are formed most often into their final shape by high-pressure extrusion, often at elevated temperature. Plastics often are joined by means of organic glues. These glues are organic polymer compounds designed to penetrate into the open, amorphous structure of plastics and break the bonds between polymer chains. When the glue dries or the glue polymer is stimulated to undergo a chemical reaction with itself and/or the material it has penetrated, the glued polymers become interlocked with polymer chains inside the surface of the softened polymers.

Metals. There are over 80 different elements that can form metals. Most of these elements can be combined with other metallic elements to form alloys. Very few metals can be found in nature with sufficient purity to be used directly. Metals often are derived from mined minerals, or ores, such as oxides, sulfides, and carbonates. Metal ores are refined or reduced to pure metal elements by a number of different reactions. Examples are: Fe2O3 + 2C → Fe + FeO + 2CO FeO + C → Fe + CO TiO2+ Cl2 + C → TiCl4 + CO2 TiCl4 + 2Mg → Ti + 2MgCl2 PbS + O2 → Pb + SO2 After refinement, the elemental metal can be further processed by combining different metals to form alloys: Fe + Cr + Ni yields stainless steel; Cu + Zn yields brass; Pb + Zn yields solder. Elements are alloyed to alter or augment specific properties—to increase corrosion resistance (i.e., Cr in stainless steel), to increase strength (i.e., Zn in brass), or to lower melting point (i.e., Pb in solder). The microstructure of the final product is controlled by the metal chemistry and the fabrication steps. During solidification, if the cooling rate is rapid, there is insufficient time for the entire system to maintain thermal equilibrium. Initially at the solid-liquid interface, there is rapid grain nucleation producing a large number of small grains. The large difference in temperature enhances preferential grain growth in certain lattice directions, resulting in the solidification front limiting the number of grains that grow and producing elongated structures, columnar grains, or elongated treelike structures, dendritic grains. As the temperature gradient is reduced, the solidification front slows, and grain nucleation can again occur resulting in equiaxed grain structure in the interior of a casting. During cold working, grains become flattened and elongated platelet grains in the rolling direction. Rapid cooling or reheating of a cold worked structure produces equiaxed grains. During the initial solidification of a metal, a high cooling rate may prevent the entire system from being in chemical equilibrium. Elemental segregation may result from different solubility levels at different temperatures and the inability of elements to diffuse rapidly through the solid to produce a uniform composition. Alloys often are held at elevated temperature after casting to allow sufficient time for atomic diffusion and to normalize, or homogenize, the composition throughout the microstructure. After melting, the metal is cast into either near-net-final shape or into slabs or billets. Slabs and/or billets are processed further by either hot or cold working. In hot working a material is deformed at elevated temperatures where the yield strength is lower than at room temperature and the material does not work harden. During cold working, often done as a final processing step after several hot working operations, the material stores up elastic energy and its yield strength is increased. The energy can be reduced by a quick heat treatment (anneal) during which the elastic energy is dissipated and recrystallization occurs, resulting in new grain development or refinement and an overall material grain size reduction. Because different elements can be combined readily to form alloys of different composition, a common means of joining alloys together is to weld them by melting their interface surface and allowing the molten material to form a new alloy during solidification. Soldering and brazing are forms of metallic glues. When applied, they are molten and the solder or braze wets (or reacts) with the metal surface, which can be thought of as a very thin weld zone. Fluxes are compounds that remove the surface metal oxide layer present on all metal surfaces and aid the solder or braze in bonding the metal surfaces. Welding melts the metal surface and can generate several changes in the microstructure of metals. The resolidified region, melt zone, and the region adjacent to the melt zone, the heat- affected zone (HAZ) will often possess a different microstructure than that of the bulk material being joined. These zones can result in changes in material properties, most often mechanical and corrosion properties. The metal near a weld is annealed and is often a region of inhomogeneity. Composites. The large range of possible structures and materials that are employed in composites (Table 2) is the result of a wide variety of processing techniques. Numerous techniques have been developed for creating particular composite structures, many of which are unique to a particular composite set of materials and microstructure. Examples include:

•

• • •

Surface processing to improve surface hardness, enhance wear resistance, increase fatigue life, and/or improve corrosion-oxidation resistance. Techniques include laser processing, ion implanting, and shot and ball peening. Surface coatings to reduce surface interactions. Techniques include thermal spray (coating with metals and ceramics), painting (coating with organics), and vapor deposition and plating. Coprocessing to place a second phase in the interior of the structure. Techniques include coextrusion of polymers to form fiberglass, metal pipes, and layering of wood between organic glues to form plywood. In situ processing of materials in which a second phase is formed as part of the processing of the material. Techniques include heat treatment of metal alloys to form precipitates (often resulting in selected grain growth characteristics) and induction heating of polymers to form a different bonding structure in the interior than on the surface.

The field of composites is expanding rapidly with an ever-expanding array of processing techniques. Care must be taken, however, as processing of the composite may change the properties of the starting materials (e.g., thermal processing of metals may eliminate the cold work structure of metals). Thought must also be given to the long-term application for which the composite is designed. Long-term exposure at elevated temperature or thermal cycling may result in debonding between the composite phases if differences in thermal expansion and thermal conductivity are too great. These differences in thermal properties must be considered, along with mechanical stress concentrations, when joining different materials together. Long-term aging of composites may result also in diffusion of elements from one composite phase into the second, changing the properties of the composite materials or forming a unique phase at the two-phase interface, again changing the microstructure and overall properties of the composite.


Material Design, Application, and Failure Materials currently are being chosen and composites designed to meet ever more demanding specifications. To ensure safe and economic designs and applications, industry, materials organizations, and government have established codes and regulations that materials are expected to meet before being placed in service. The codes or regulations require extensive testing before the material receives certification. These requirements are based on experience and knowledge learned over years of trial and error. Material failure can occur because the wrong material was used, the material had a preexisting manufacturing defect, there was an engineering design flaw, abuse occurred and the material was degraded, or the design load limits were exceeded. With the development of better understanding of materials and their responses to different environments, the choice of a material for a particular application is no longer by chance or limited to past experience. Manufacturing techniques have improved significantly, and reproducibility and reliability of a consistent final product is realized by implementation of quality assurance programs. Engineering design has been improved greatly by computer design programs that allow for ever greater analysis. Current software programs can be used to understand the effect of environment and catastrophic events, to incorporate safety factors, to determine when safety checks should be performed, and to predict service lifetime before failure. All products eventually fail or become no longer usable or obsolete. The service life of materials can be extended by proper maintenance and inspection. Proper material design and understanding can produce a product that performs safely for a specified time before it needs to be taken out or service; that is, before loss of function, catastrophic failure, and damage occurs. Most of the failures of today are not the result of a single

factor but result from a number of one-time events, none of which would cause failure by itself but when all are taken together, result in failure. The more common events that lead to failure are: • • • •

Failure to note and report the occurrence of an unusual event (such as an overload of one type or another) The failure of inspection (either not performed or incomplete) Change of procedure or application Change of environment

The study of materials is a continuously evolving field. New demands are placing greater emphasis on new applications of old materials, on newer materials with better properties, and on greater understanding of both the fundamentals of material science and engineering design.


Selected References General • • • • • • • • • • •

Annual Book of ASTM Standards, ASTM International M.F. Ashby and D. Jones, Engineering Materials, 2nd ed., Vol 1 and 2, Butterworth-Heinemann, 1996 M.F. Ashby, Materials Selection in Mechanical Design, 2nd ed., Butterworth-Heinemann, 1999 D.R. Askeland, The Science and Engineering of Materials, 3rd ed., PWS Publishing, 1999 ASM Handbook, Vol 20, Materials Selection and Design, ASM International, 1997 G.S. Brady, H.S. Clauser, and J.A. Vaccari, Materials Handbook, 15th ed., McGraw-Hill, 2002 K.G. Budinski and M.K. Budinski, Engineering Materials: Properties and Selection, 7th ed., Prentice Hall, 2001 T.H. Courtney and D. Simchi-Levi, Mechanical Behavior of Materials, 2nd ed., McGraw- Hill, 1999 H.E. Davis, G. Houck, and G.E. Troxell, Testing of Engineering Materials, 4th ed., McGraw-Hill, 1982 Engineered Materials Handbook Desk Edition, ASM International, 1995 D. Henkel and A.W. Pense, Structures and Properties of Engineering Materials, 5th ed., McGraw-Hill, 2001

Metals • • • • • • • • •

Properties and Selection: Irons, Steels, and High-Performance Alloys, Vol 1, ASM Handbook, ASM International, 1990 Properties and Selection: Nonferrous Alloys and Special-Purpose Materials, Vol 2, ASM Handbook, ASM International, 1990 H. Chandler, Metallurgy for the Non-Metallurgist, ASM International, 1998 J.R. Davis, Ed., Metals Handbook Desk Edition, 2nd ed., ASM International, 1998 G.E. Dieter, Mechanical Metallurgy, 3rd ed., McGraw-Hill, 1986 R.W. Hertzberg, Deformation and Fracture Mechanics of Engineering Materials, John Wiley and Sons, 4th ed., 1995 R.E. Reed-Hill and R. Abbaschian, Physical Metallurgy Principles, 3rd ed., PWS Publishing, 1992 L. Van Vlack, Elements of Materials Science, 6th ed., Addison-Wesley, 1989 J.R. Brown, Ed., Foseco Foundryman's Handbook, 10th ed., Butterworth-Heinemann, 1994

Ceramics • • • • • •

M.W. Barsoum, Fundamentals of Ceramics, McGraw-Hill, 1996 Ceramics and Glasses, Vol 4, Engineered Materials Handbook, ASM International, 1991 W.D. Kingery, H.K. Bowen, and D.R. Uhlman, Introduction to Ceramics, 2nd ed., John Wiley & Sons, 1976 S. Mindess, J.F. Young, and D. Darwing, Concrete, 2nd ed., Prentice Hall, 2002 J.S. Reed, Principles of Ceramic Processing, 2nd ed., John Wiley and Sons, 1995 J.B. Wachtman, Structural Ceramics, Academic Press, 1989

Polymers • • • • • •

Engineered Materials Handbook, Vol 2, Engineering Plastics, ASM International, 1988 J.R. Fried, Polymer Science and Technology, 2nd ed., Prentice-Hall, 2003 C.A. Harper, Handbook of Plastics, Elastomers, and Composites, 4th ed., McGraw-Hill, 2002 Modern Plastics Encyclopedia, McGraw-Hill (annual) A.B. Strong, Plastics—Materials and Processing, Prentice Hall, 2000 Z. Tadmor and C.G. Gogos, Principles of Polymer Processing, John Wiley & Sons, 1979

Wood • •

H.E. Desch, Timber: Structure, Properties, Conversion, and Use, 7th ed., Haworth Press, 1996 B. Madsen, Structural Behaviour of Timber, Timber Engineering Ltd., 1992

Composites • • •

Composites, Vol 21, ASM Handbook, ASM International, 2001 C.T. Herakovich, Mechanics of Fibrous Composites, John Wiley and Sons, 1997 S. Suresh, A. Mortensen, and A. Needleman, Fundamentals of Metal-Matrix Composites, ButterworthHeinemann, 1993

Processing • • • • • • • • • • • • •

Alloy Phase Diagrams, Vol 3, ASM Handbook, ASM International, 1992 Heat Treating, Vol 4, ASM Handbook, ASM International, 1993 Surface Engineering, Vol 5, ASM Handbook, ASM International, 1995 Welding, Brazing, and Soldering, Vol 6, ASM Handbook, ASM International, 1993 Powder Metal Technologies and Applications, Vol 7, ASM Handbook, ASM International, 1998 Forming and Forging, Vol 14, ASM Handbook, ASM International, 1988 Casting, Vol 15, ASM Handbook, ASM International, 1988 Machining, Vol 16, ASM Handbook, ASM International, 1989 C.R. Brooks, Heat Treatment, Structure, and Properties of Nonferrous Alloys, ASM International, 1990 J.W. Christian, The Theory of Transformations in Metals and Alloys, 2nd ed., Pergamon Press, 1975 G. Krauss, Steels: Heat Treatment and Processing Principles, ASM International, 1990 J.W. Martin, Precipitation Hardening, 2nd ed., Butterworth-Heinemann, 1999 H. Okamoto, Desk Handbook: Phase Diagrams for Binary Alloys, ASM International, 2000

Testing and Failure Analysis • • • • • •

Mechanical Testing and Evaluation, Vol 8, ASM Handbook, ASM International, 2000 Metallography and Microstructures, Vol 9, ASM Handbook, ASM International, 1985 Materials Characterization, Vol 10, ASM Handbook, ASM International, 1986 Failure Analysis and Prevention, Vol 11, ASM Handbook, ASM International, 2002 Fractography, Vol 12, ASM Handbook, ASM International, 1987 Nondestructive Evaluation and Quality Control, Vol 17, ASM Handbook, ASM International, 1989

• • • • • • •

Friction, Lubrication, and Wear Technology, Vol 18, ASM Handbook, ASM International, 1992 Fatigue and Fracture, Vol 19, ASM Handbook, ASM International, 1996 V.J. Colangelo and F.A. Heiser, Analysis of Metallurgical Failures, 2nd ed., John Wiley & Sons, 1987 J.A. Collins, Failure of Materials in Mechanical Design, 2nd ed., John Wiley & Sons, 1993 K.A. Esaklul, Handbook of Case Histories in Failure Analysis, ASM International, Vol 1, 1992; Vol 2, 1993 A.J. McEvily, Metal Failures: Mechanisms, Analysis, Prevention, John Wiley & Sons, 2002 D.J. Wulpi, Understanding How Components Fail, 2nd ed., ASM International, 1999

S. Shah, Applications of Modern Analytical Instruments in Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 992–998

Applications of Modern Analytical Instruments in Corrosion Sadiq Shah, Western Illinois University

Introduction CORROSION is a redox process at the metal- liquid or metal-air reactive interface. Identifying effective means to prevent corrosion requires an understanding of corrosion mechanisms under a variety of environmental conditions. However, unraveling corrosion mechanisms is not always a straightforward process. An understanding of the mechanistic details requires conducting experiments under simulated and controlled laboratory conditions. Ideally, it requires real time monitoring of the corrosion reactions under these conditions at a molecular or atomic level. To gain an insight into the mechanism of corrosion, both in situ (at the original location) and ex situ (away from the original location) techniques have been employed successfully over the last couple of decades (Ref 1). However, real time in situ experiments generally require controlled laboratory conditions. Several articles in the Section “Corrosion Testing and Evaluation” address controlled laboratory testing. Alternatively, one can use ex situ analytical techniques that are capable of analyzing surfaces before and after the corrosion reaction, thus following the changes in metallurgy and chemistry at the interfaces under a given set of environmental conditions. This article reviews a variety of surface-sensitive ex situ analytical techniques that have been successfully used in understanding corrosion processes and identifying the nature of protective corrosion inhibitor films. The focus of this review is to explain the principles of various surface-sensitive techniques and the usefulness and limitations of these techniques (Ref 2, 3, 4, 5). To gain additional insight into the use of various surface-sensitive analytical techniques for applications in corrosion science, various examples from the recent literature are cited.

References cited in this section 1. S. Shah, Ed., Surface and Interface Characterization in Corrosion, NACE International, 1994 2. M. Riviere, Handbook of Surface and Interface Analysis Methods of Problem Solving, Marcel Dekker, Inc., 1998 3. R. Howland, L. Benatar, and C. Symanski, A Practical Guide to Scanning Probe Microscopy, DIANE Publishing Co., Collingdale, PA, 1998

4. J. Lipkowski and P.N. Ross, Imaging of Surfaces and Interfaces, John Wiley & Sons, Inc., 1999 5. A. Hubbard, The Handbook of Surface Imaging and Visualization, CRC Press, 1995

S. Shah, Applications of Modern Analytical Instruments in Corrosion, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 992–998 Applications of Modern Analytical Instruments in Corrosion Sadiq Shah, Western Illinois University

Strategy for Selecting Analytical Approaches Various techniques can be used to analyze corrosion product samples removed from the surface and also small samples of the base metallurgy. However, this approach may require dissolving the sample in a suitable solvent system and then analyzing the solution. For such samples, one needs to be aware of the limitations during the sample preparation step. These limitations include the possibility of a change in the chemical nature of the corrosion product or incomplete sample transfer during sample preparation steps. This can result in revealing inaccurate or incomplete information about the chemical composition of the sample. An ideal approach is to analyze surfaces and corrosion-product deposits by direct surface-sensitive analytical techniques when possible. It should be pointed out that no single analytical technique can answer all questions well; instead, a strategy involving the use of a suitable combination of analytical techniques is necessary to learn as much about the corrosion mechanism as possible. The Selected References cited here clearly demonstrate the importance of this strategy in corrosion science. A clear understanding of the mechanism of corrosion will help identify effective corrosion analysis and prevention techniques. Modern analytical methods for analyzing surfaces for corrosion and corrosion inhibition processes as well as for failure analysis can be classified into two main categories: (a) surface structure and (b) chemical identity and composition. These categories are based strictly on the nature of the questions to be addressed, such as insight into surface topography or surface structure, or understanding the chemical nature and identity at a molecular and atomic level. Table 1 lists a broad spectrum of the surface-sensitive analytical techniques, the principles of respective techniques, and the information they reveal. These techniques are described in more detail under the section devoted to each analytical method. Table 1 Surface sensitive analytical techniques, principles and applications Analytical technique Surface structure techniques Scanning electron microscopy (SEM)

Principle

Incident electron beam generates a secondary electron emission that is used to generate the surface image. Microscopic force sensor (cantilever) is Atomic force microscopy used to sense the force between a sharp tip (AFM) and the sample surface as the sample is scanned to generate an image. Tunneling current is monitored as the probe Scanning tunneling tip is scanned over a surface of interest in microscopy (STM) the x-y plane to generate an image. Reflected light is used to generate a Optical microscopy magnified image. Chemical identity and composition techniques

Target information Microscopic imaging of the surface structure Imaging of insulated surface structure at atomic resolution

Imaging of conducting surface structure at atomic resolution Macroscopic surface structure details

Energy dispersive x-ray spectroscopy (EDS) Wavelength dispersive xray spectroscopy (WDS)

X-ray diffraction (XRD)

Auger electron spectroscopy (AES)

X-ray photoelectron spectroscopy (XPS) or ultraviolet photoelectron spectroscopy (UPS) Ion scattering spectroscopy (ISS) Secondary ion mass spectrometry (SIMS) Extended x-ray absorption fine structure (EXAFS)

Fourier transform infrared absorption spectroscopy (FTIR) Raman and surface enhanced Raman spectroscopy (Raman, SERS)

Incident electron beam generates emission of x-rays characteristic of elements present at the surface. The principle is the same as that for EDS except the emitted x-rays are measured one wavelength at a time. Diffraction of the incident x-ray beam from various plans of crystal lattice create a diffraction pattern characteristic of the sample. Incident electron beam initiates a multistep process to facilitate the ejection of an outer shell electron. The energy of this ejected electron is characteristic of the surface atoms. Incident x-rays or ultraviolet photons on the surface eject photoelectrons. The energy of the photoelectrons is characteristic of the surface atoms. The energy of scattered primary ions from the surface allows identification of surface atoms. Incident ion beam ejects the surface atoms as ions, and mass of these secondary ions is measured. The x-ray absorption generates the interference effects between emitted photoelectron waves and backscattering waves characteristic of the local structure. The absorption of the infrared photons results in vibrational excitation that is characteristic of the surface molecules and the environment. Energy shifts of the scattered photons are characteristic of the molecular identity.

Elemental identification of the surface species Elemental identification of the surface species with higher resolution and sensitivity than EDS Elemental and phase identity and composition of inorganic corrosion product Elemental identity and composition of the surface species and the depth profile

Elemental identity and composition of the surface species, the electronic structure, chemical bonding, and elemental depth profile Chemical composition of the surface films at atomic and a molecular level Chemical composition of the surface adsorbed species Composition of adsorbed species, number, and separation distances of surface atoms Vibrational structure of molecules and bonding interactions with the surface and the surroundings Vibrational structure of adsorbed molecules on the surface


Surface-Structure Analysis Examination of the surface structure pre- and post-corrosion can often reveal clues to the nature of the corrosion process, the role of the base metal, and the role of environmental conditions in the process. For

surface topographic details one may utilize microscopic techniques capable of revealing surface structure from the macroscopic level to the atomic level. These techniques include optical microscopy, scanning electron microscopy (SEM), atomic force microscopy (AFM), and scanning tunneling microscopy (STM) (Ref 6). Optical microscopes can magnify the surface structure to a certain magnification level. Higher magnification to micron levels requires the use of SEM, and for atomic and molecular level resolution, AFM or STM is required. While both AFM and STM are capable of revealing the structure at atomic and molecular levels, AFM is generally useful for electrically nonconducting surfaces and STM for conducting surfaces. When considering the use of SEM on organic surfaces, the analyst must foresee the possibility of thermal damage to the surface.

Reference cited in this section 6. L.A. Bottomley, J.E. Coury, and P.N. First, Scanning Microscopy, Anal. Chem., Vol 68, 1996, p 185R– 230R


Determination of Chemical Composition An understanding of the chemical identity of a metal surface before and after corrosion can reveal the mechanistic details of the corrosion reaction. The study may require analysis of the corrosion products or the corrosion inhibitor films after the corrosion reaction. Compositional analysis can also suggest the nature of the interaction between the metal surface and the aggressive species that led to the corrosion reaction or interaction of the corrosion inhibitor molecules for corrosion protection (Ref 7). A variety of analytical tools can be used to gain such information. Thus the analytical techniques for surface analysis can be divided into two classes, namely, (a) techniques used for the analysis of the metal surface and metal oxides and (b) those used for the analysis of the corrosion product and corrosion protective organic films. Analysis of Metal Surfaces and Metal Oxides. The analysis of the metal samples to identify the chemical nature of the base metallurgy and the surface oxide films, including the protective passive film, can be accomplished by the use of various surface sensitive analytical techniques. These techniques include energy dispersive x-ray spectroscopy (EDS), wavelength dispersive x-ray spectroscopy (WDS), Auger electron spectroscopy (AES), xray photoelectron spectroscopy (XPS) also referred to as electron spectroscopy for chemical analysis (ESCA), ultraviolet photoelectron spectroscopy (UPS), ion scattering spectroscopy (ISS), secondary ion mass spectrometry (SIMS) and extended x-ray absorption fine structure (EXAFS). While x-ray diffraction (XRD) is not a surface technique, it is widely used for the analysis of metals, alloys, and the mineral contents and phase identifications of corrosion product. Analysis of Corrosion Products and Protective Inhibitor Films. For the analysis of the corrosion product or protective films that are expected to be primarily inorganic in nature, EDS, AES, XPS, ISS, SIMS and EXAFS generally are suitable techniques. However, for corrosion inhibitor films that are primarily organic in nature, Fourier transform infrared (FTIR) spectroscopy, Raman or surface enhanced Raman spectroscopy (SERS), SIMS, ISS, and XPS are the most appropriate techniques. For routine analysis of the molecular nature of the film, FTIR is the most versatile and inexpensive nonvacuum technique. Even though Raman and SERS can also provide information regarding the molecular nature of the film, these techniques require a fair amount of additional instrumental sophistication and experience. The vacuum techniques SIMS, ISS, and XPS require expensive instrumentation and extensive experience. A number of private analytical service laboratories around the country offer these techniques.

Reference cited in this section 7. S. Shah, Corrosion: An Interfacial Process, Corrosion 91, “Paper 75,” 1991, p 1–22


Analytical Methods In general, surface-sensitive techniques involve irradiation of the surface of interest with photons, x-rays, electrons, or an ion beam. The scattered, emitted, or reflected beam then carries with it information regarding the chemical identity of the surface atoms or molecules. An overview of commonly used surface-sensitive techniques for analysis of the surface, atomic, and molecular structures is given next.

Surface Topological Structure Surface structural details can be examined by a variety of the analytical techniques with resolution ranging from the macroscopic level to the atomic and molecular level. These techniques are based on varied principles. Optical Microscopy. A simple optical microscope can magnify the surface topological details to a level revealing critical details not visible to the naked eye. With the use of appropriate optics, the microscope utilizes optical reflections from a surface of interest to generate an image of the surface structure. However, it does not allow for the identification of the chemical nature of surface structures. The spatial resolution of this technique is limited to approximately 1 μm. This simple technique generally can be an important first step to survey the macroscopic surface details of a sample. Scanning Electron Microscopy. In contrast to optical microscopy, which relies on photons, the SEM approach relies on an incident beam of electrons. Under ultrahigh vacuum an electron beam is focused on the surface of interest, and the secondary electrons emitted from the surface are used to generate the image of the surface structural details. The incident beam energy can be as high as 30 keV (1.15 kcal); however, the secondary emitted electrons have energy of a few tens of eV (cal). The images generated by the emitted electrons are sensitive to the orientation of the incident beam, the relative orientation of the detector, and the organic or inorganic chemical nature of the surface. The spatial resolution is on the order of less than 100 nm (3.9 × 10-6 in.). SEM can magnify surface structures from 10 to 30,000 times. Care should be exercised in examining organic surface films or polymeric samples. These organic films can be thermally degraded by the high-energy incident electron beam. Thus, the lowest possible energy should be used for the incident electron beam. To improve the contrast, some surfaces may require gold, carbon, or palladium coating prior to SEM examination; however, here too thermal degradation must be a consideration. This technique has been used routinely for analyzing the surface structure, including the failure analysis of structures used in a broad range of applications (Ref 8, 9, 10, 11, 12, 13, 14). While SEM imaging relies on the emission of secondary electrons from the sample surface, another similar technique relies on the back-scattered incident electrons and is referred to as back-scattered electron imaging (BEI). In this technique the intensity of the back-scattered electrons depends on the atomic number of the elements present in the sample. The higher the atomic number of the elements present in the sample, the higher is the number of the back-scattered electrons. As a result, lower atomic number elements absorb more incident electrons and therefore appear dark in the micrographs. Similarly, higher atomic number elements scatter a larger number of the incident electrons, thereby making the micrograph bright. Thus on flat samples this technique can be

informative. In general this sensitivity of BEI to the atomic number can be used to make qualitative decisions about structure and compositions. An extension of SEM is environmental scanning electron microscopy (E-SEM) where samples can be examined under relatively higher pressures than those involved in the traditional SEM (Ref 15). The pressure conditions allow for examination of the surface structure under atmospheres of water vapors, air, argon, nitrogen, or other gases or vapors. This allows for a dynamic characterization of the corrosion process and wetting, drying, melting, and crystallization processes. However, the instrumentation is not as common as that used for the traditional SEM. Scanning Tunneling Microscopy (STM). In STM, a small bias voltage is applied between the probe tip and the sample, and the tunneling current is monitored as the probe is scanned in the x-y plane over the surface of interest. The probe tip is typically composed of platinum. The surface structure can be evaluated under a variety of environmental conditions such as ambient conditions, ultrahigh vacuum (UHV), air, cryogenic fluids, and liquid solutions. STM is capable of monitoring the surface structural details in situ, including the progression of pitting (Ref 16, 17, 18, 19, 20, 21). Samples can be conducting and semiconducting in nature. Atomic force microscopy (AFM) is the counterpart to STM and is also used to examine surface structure (Ref 22, 23, 24, 25, 26). It utilizes a force sensor called a cantilever to sense the force between a sharp tip and the sample surface as the sample is scanned in the x-y direction. The signal is processed to generate the surface image. The AFM technique allows the examination of the surface structure of insulating substrates.

Atomic and Molecular Structure Energy Dispersive X-Ray Spectroscopy (EDS). This technique is used in conjunction with a scanning electron microscope (SEM). During the SEM examination any surface structure of interest can be analyzed for elemental identity and elemental composition; the EDS technique can identify elements with atomic numbers greater than 5 (boron). The principle of the technique involves the analysis of the energy and the intensity of the emitted x-rays as the incident electron beam is focused on the surface of interest. The incident electron beam ejects an electron from an inner shell; an electron from a higher energy level then fills the vacancy generated by the ejected electron. The electron filling the vacant shell loses energy corresponding to the difference between the two energy levels involved in the form of x-rays. The energy of the emitted x-rays is characteristic of the elements involved (Fig. 1b).

Fig. 1 Energy-level diagrams showing the electron transitions that form the basis for three techniques. (a) X-ray photoelectron spectroscopy (XPS). (b) X-ray analysis. (c) AES (Auger electron spectroscopy). φ, spectrometer work function The instrument generates a spectrum of the relative intensity of the x-rays versus the energy of the x-rays, thereby allowing for the identification of various elements with atomic numbers greater than 5. The convenience of examining the structural details through SEM with the ability to focus on any structure of interest to gain information about the elemental identity of this structure becomes a powerful approach. The technique is available widely and relatively easy to use by trained professionals. With the use of known standards, the technique can be used for quantitative analysis of elements; however, it is commonly used for

qualitative examination of samples. Nevertheless, the EDS spectrum does reveal the relative concentration differences for various elements across the surface or between various samples compared. This technique is very useful for the analysis of surface contamination, corrosion products, preliminary identification of an alloy or surface coating, and phase identification (Ref 27, 28, 29, 30). The combination of SEM and EDS can also be used for failure analysis. Wavelength dispersive x-ray spectroscopy (WDS) is another x-ray technique that can be used for the identification of the elemental composition of a sample of interest. In principle, WDS is a technique complementary to EDS. The main differences between these two techniques are that in EDS, the x-ray photons are evaluated for all energies at once and displayed against energy, and in WDS, x-ray photons are evaluated based on their wavelengths with a single wavelength measured at a time. Typically, WDS offers better resolution and sensitivity than EDS; however, the inherent nature of the slow measurement process makes this a slower technique than EDS. While not necessarily a surface technique, x-ray diffraction is widely used and is a valuable technique for determining mineral identities of surface films and corrosion products that are suitably thick. Auger electron spectroscopy (AES) involves the use of an incident beam of electrons under ultrahigh vacuum (UHV) (Ref 31). The incident beam is focused on the surface of interest, and the energy of the emitted beam is analyzed. In AES spectroscopy the incident beam ejects an inner electron. Figure 1(c) illustrates the case of ejection of the core electron (K-shell). The vacancy created may be filled by an electron from a higher energy shell, in this case the L-shell. The energy corresponding to the difference of the K and L shells is then consumed in expelling an outer electron (Auger electron) from the atom (Fig. 1c). Auger electrons have low energies, and therefore only electrons with sufficient energy originating from the surface or near the surface from a depth of approximately 0.8 to 2 nm (3.1 to 7.9 × 10-8 in.) can escape. Auger electron spectroscopy has been used for not only elemental identity of the surface but also for determining the depth profiles of elements such as the composition of oxide films including the passive films and organic corrosion inhibitor films (Ref 32, 33, 34, 35). Here again quantitative analysis can be conducted with the use of appropriate standards; however, use of a semiquantitative approach is more common. X-ray photoelectron spectroscopy (XPS), also referred to as electron spectroscopy for chemical analysis (ESCA), is very similar to AES, except that an incident monochromatic x- ray beam is used and the energy of the photo- ejected electrons from the surface atoms or molecules is analyzed (Fig. 1a). These photo-ejected electrons from the surface carry information regarding the chemical identity and environment of the surface atoms and molecules. Therefore, the peak position of the photo-ejected electrons in the XPS spectrum depends on not only the atomic number but also the chemical environment of an atom or molecule. Thus, information regarding the functional groups adsorbed on the surface can be determined. Since the soft x-rays are less damaging, insulators with elements of atomic number greater than 4 (beryllium) can be analyzed to a depth of about 1 to 10 nm (3.9 × 10-8 to 3.9 × 10-7 in.). The technique can be used for quantitative analysis with the use of appropriate standards. X-ray photoelectron spectroscopy has been extensively used for the depth profiles of the elemental composition of surface films, both organic and inorganic in nature (Ref 36, 37, 38, 39, 40, 41, 42, 43). Such information allows for not only the chemical nature of the surface films but also the mechanistic details of the film forming processes. Ion Scattering Spectroscopy. This technique involves scattering ions from the surface of interest. The measurement of the energy of the scattered ions allows for determining the chemical identity of the surface or the surface film (Ref 44, 45, 46, 47). The technique utilizes noble gas ions such as 3He+, 4He+, 20Ne+, or 40Ar+ at energies between 0.5 to 5.0 keV (1.91 × 10-2 to 1.91 × 10-1 kcal). Generally, a target ion mass less than the probe incident ion mass will not be detected. This technique has been used to examine surface adsorbed films, surface oxides, and corrosion inhibitors. The depth resolution can be at an atomic level. A major limitation for quantitative analysis with ISS is the neutralization of the scattered ions. Secondary Ion Mass Spectrometry. The principle of this technique involves the use of an energetic incident ion beam to eject secondary ions from the surface of interest (Ref 48, 49, 50, 51). The ejected ions are analyzed by mass spectrometry to determine the chemical identity. The high yield of the secondary ions is a direct result of the high flux of the incident ion beam and the ionization probability. Thus, the depth profile of the elemental composition of the surface films can be easily monitored. The use of efficient ion collectors in the detection mode allows for good sensitivity. The SIMS technique has been used for analysis of surface adsorbed species, passive films, and surface contaminants. The depth resolution is at a molecular level. The analyses include surface oxides, hydroxides, and corrosion inhibitor chemistries. For organic corrosion inhibitor films, one must

be aware of the possibility of the surface charging. An advantage of this technique is that the detection of surface ions or clusters of ions allows for the identification of the molecular nature of the organic and inorganic protective films. This information provides an insight into the mechanistic details of the formation of the protective film. Reflectance Fourier Transform Infrared Absorption Spectroscopy. Fourier transform infrared absorption spectroscopy (FTIR) relies on the vibrational excitation of the surface molecules (Ref 52). The infrared incident photons are absorbed by the appropriate functional groups adsorbed on a surface. The intensity of the reflected photons as a function of the energy of the photons results in a fingerprint spectrum of organic and some inorganic species. This allows for the identification of the chemical nature of the film at a molecular level (Ref 53, 54, 55, 56, 57, 58, 59). This technique is primarily used for organic surface films; however, it can be used for some inorganic surface films as well. The technique is a relatively inexpensive approach for the analysis of relatively thick films of organic corrosion inhibitors or polymeric coatings on surfaces. However, extreme care should be taken in analyzing very thin films. FTIR can be used in ex situ and in situ modes. The ex situ mode allows for a relatively easy characterization of the surface films (Ref 60). Attenuated total reflectance (ATR) is a sampling mode where the sample of interest is pressed against a flat window. Multiple infrared beam reflections from the surface of the film through the window carry the functional group information. Another sampling mode uses multiple reflections within the film (not the sampling window), such as crystalline materials, resulting in a diffuse reflectance infrared Fourier transform (DRIFT) spectrum. The DRIFT spectrum generally contains some specular reflectance that can potentially distort the spectral resolution. Attenuated total reflectance is the most commonly used sampling mode. The windows are generally composed of ZnSe, Si, or Ge. The ZnSe windows are attacked by strong acids, but Si and Ge are very inert. Each material has its own characteristic spectral operating range. Thus materials must be appropriately selected for the specific application of interest. Raman and Surface Enhanced Raman Spectroscopy. Raman spectroscopy evaluates the shifts in the energy of the incident visible and ultraviolet photons at right angle to the incident beam. This measurement allows for the identification of the vibrational structure of the surface adsorbates, hence the molecular nature of the species (Ref 61, 62, 63, 64, 65, 66, 67, 68). The Raman scattering signal is generally weak; therefore an intense source of light such as a laser is needed for these measurements. The incident beam is typically an Ar+ or a Kr+ laser. For SERS the choice of the laser frequency that coincides with the electronic structure of the species of interest allows for thousand-fold signal enhancement. Thus SERS becomes an advantageous technique compared to Raman spectroscopy for some samples of interest. Surface enhanced Raman spectroscopy signal is typically observed for gold, copper, and silver. For SERS inactive steel, surfaces are typically coated with silver islands to make the surface SERS active. Both of the preceding techniques have been successfully used for examining the molecular nature of passive films and corrosion inhibitor films. Corrosion related Raman or SERS studies have been carried out on a broad range of alloys and protective corrosion inhibitor films in both in situ and ex situ modes (Ref 69, 70, 71, 72, 73, 74, 75, 76). Raman spectroscopic technique also has some limitations that need to be considered for various samples of interest. For dark and colored samples, the high absorption coefficient of the incident visible light limits the strength of the scattered Raman signal. Although not common, in some instances any fluorescence signal can complicate the spectrum, thus making the measurements difficult. To improve the Raman signal, intense incident laser light is used, and it sometimes presents a challenge of thermal degradation of the sample of interest. Extreme care is thus needed in conducting experiments to avoid chemical compositional changes. In general, these limitations can be addressed by appropriate experimental precautions. Extended X-Ray Absorption Fine Structure. X-ray absorption spectroscopy is used for a number of applications (Ref 77). This is primarily due to the availability of synchrotron radiation sources that can generate x-rays with intensities that are significantly higher than conventional x- ray sources. The absorption of x-rays by materials generally decreases as the energy of the incident x-ray beam is increased. However, when the energy of the xray reaches a threshold to promote a core electron to an unoccupied valence level or the continuum, absorption suddenly increases. The structure of this absorption edge is referred to as the x-ray absorption near edge structure (XANES). The energy of the x- ray edge corresponding to this absorption is characteristic of the individual elements, thus allowing for the identification of specific elements (Ref 78, 79, 80). Extending beyond the XANES region, from 100 to 1000 to 2000 eV (3.8 to 38.3 to 76.6 cal) above the edge are oscillations that are referred to as the extended x-ray absorption fine structure (EXAFS) (Ref 81, 82, 83, 84, 85,

86, 87, 88, 89, 90, 91). The EXAFS results from the oscillatory modulation of the x-ray absorption cross section by the interference between the emitted photoelectron waves and the backscattered waves from neighboring atoms in the molecule. The EXAFS allows for determining the interatomic distances of the nearest neighbors of a given atom and therefore the molecular structure of the surface film. The technique can be used in situ and ex situ to monitor electrochemical reactions. The technique has been used successfully to evaluate the chemical nature of conversion coatings (Ref 92). This technique does require access to the synchrotron radiation source and sufficient theoretical and experimental experience. More detailed descriptions of the principles and applications of these techniques are provided in Materials Characterization, Volume 10 of ASM Handbook, and in the section “Atomic Force Microscopy” in the article “Measurement of Surface Forces and Adhesion” in Friction, Lubrication, and Wear Technology, Volume 18 of ASM Handbook.

References cited in this section 8. T.G. Taranova, G.M. Grigorenko, V.A. Kostin, and V.V. Arsenyk, Applications of Scanning Electron Microscopy in Modern Material Science, Welding and Related Technologies, Welding International, Vol 14 (No. 9), 2000, p 745–748 9. Y. Ding and D.O. Northwood, Methods For Preparation of Cross-Sectional Scanning Electron Microscopic Specimens and Their Applications to Corroded Specimens of a Zirconium Alloy and Titanium Nitrogen Coated Stainless Steel, Mater. Charact., Vol 29 (No. 1), 1992, p 25–33 10. S.C. Tjong, Electron Microscope and Raman Characterization of The Surface Oxides Formed on The Fe-Cr Alloys at 400–800°C, Mater. Charact., Vol 26 (No. 1), 1991, p 29–44 11. R.K.R. Singh, A.K. Tyagi, K. Krishan, and J.B. Gnanamoorthy, Secondary Ion Mass Spectrometery and Scanning Electron Microscopic Characterization of Grain Boundary Oxide Ridges in 9Cr-1Mo Steels Having Different Silicon Contents, and Influence of Grain Size on Scale Spalling, Mater. Sci. Technol., Vol 10 (No. 7), 1994, p 592–598 12. L. Veleva and M. Luja, Scanning Electron Microscopic Characterization of Copper Corrosion Products, Br. Corros. J., Vol 34 (No. 1), 1999, p 34–36 13. S. Shah and M. Bernardo, Reusable Surgical Instruments: Metallurgy and Corrosion, Medical Product Sales, Vol 31 (No. 4), 2000, p 26–27 and Vol 31 (No. 6), 2000, p 26 14. P.J.R. Uwins, Environmental Scanning Electron Microscopy, Mater. Forum, Vol 18, 1994, p 51–57 15. K.P.J. Williams, I.C. Wilcock, I.P. Hayward, and A. Whitley, Applications of State-of- the-Art Raman Microscopy and Direct Raman Imaging. Progress in Corrosion, Mineralogy, and Materials Research, Spectroscopy, Vol 11 (No. 3), 1996, p 45–50 16. J. Kunze, V. Maurice, H.L. Klein, H.H. Strehblow, and P. Marcus, In Situ Scanning Tunneling Microscopic Study of The Anodic Oxidation of Copper (111) in 0.1 M NaOH, J. Phys. Chem. B, Vol 105 (No. 19), 2001, p 4263–4269 17. V. Maurice, H.H. Strehblow, and P. Maurice., In Situ STM Study of the Initial Stages of Oxidation of Cu (111) in Aqueous Solution, Surf. Sci., Vol 458 (No. 1–3), 2000, p 185–194 18. E.P. Friis, J.E.T. Anderson, L.L. Madsen, P. Moller, and J. Ulstrup, In Situ STM and AFM of The Copper Protein Pseudomonas aeruginosa azurin, J. Electroanal. Chem., Vol 431 (No. 1), 1997, p 35–38 19. D.J. Noll, P.G. Vanpatten, M.A. Nicholson, K. Booksh, and M.L. Myrick, Flow-Injection System for The Scanning Tunneling Microscope, Rev. Sci. Instrum., Vol 66 (No. 8), 1995, p 4150–4156

20. G.P. Bierwagen, R. Twite, G. Chen, and D.E. Tallman, Atomic Force Microscopy, Scanning Tunneling Microscopy and Electrochemical Characterization of Al Alloys, Conversion Coatings, and Primers Use for Aircraft, Prog. Org. Coatings, Vol 32 (No. 1–4), 1997, p 25–30 21. F.P. Zamborini, J.K. Campbell, and R.M. Crooks, Spectroscopic, Voltametric, and Electrochemical Scanning Tunneling Microscopic Study of Underpotentially Deposited Cu Corrosion and Passivation With Self-Assembled Organomercaptan Monolayers, Langmuir, Vol 14 (No. 3), 1998, p 640–647 22. C.J. Rubim, H. Jae, E. Henderson, and T.M. Cotton, Surface Enhanced Raman Scattering and Atomic Force Microscopy of Brass Electrodes in Sulfuric Acid Solution Containing Benzotriazole and Chloride Ion, Appl. Spectrosc., Vol 47 (No. 1), 1993, p 80–84 23. W. Kautek, M. Geu, M. Sabre, and P. Ahao, Multi-Method Analysis of Metal/Electrolyte Interface: Scanning Force Microscopy, Quartz Microbalance Measurements, FTIR and Grazing Incidence X-ray Diffractometry at Polycrystalline Electrode, Surf. Interface Anal., Vol 25 (No. 7–8), 1997, p 548–560 24. R.K.R. Singh and A.K. Tyagi, Secondary Ion Mass Spectrometry and Scanning Electron Microscopic Analysis of Grain Boundary Oxides in 9Cr-1Mo Steel and Influence of Grain Size on Scale Spalling, Mater. Sci. Technol., Vol 10 (No. 1), 1994, p 27–34 25. P. Lemoine and J.M. McLaughlin, Continuity and Topography of Ultrathin Diamond-Like Carbon Films Characterized by Scanning Electron Microscopy, Energy Dispersive X-ray Analysis and Atomic Force Microscopy, J. Vac. Sci. Technol., Vol 17 (No. 1), 1999, p 176–182 26. M. Bethencourt, F.J. Botana, J.J. Calvino, M. Marcos, J. Perez, and M. and A.R. Chacon, A SEM and EDS Insight Into The Corrosion Process of AA 5083 Al-Mg Alloy in NaCl Solutions, 14th Electron Microscopy Proc. of International Congress, B. Calderon, A. Hector, and M.Y. Jose, Ed. Vol 2 (No. 103–104), Institute of Physics Publishing, 1988 27. K. Forkel, D. Born, F.G. Wihsmann, and R. Stodolski, A SEM/EDXA Study on the Chemical Attack of Aqueous NaOH/KOH Solutions on the Surface of Metal-Metalloid Glass (Fe,Cr)80(P,C,Si)20 and Fe75Cr5P8C10SiO2, Respectively, Mikrochim. Acta, Vol 125 (No. 1–4), 1997, p 407–413 28. G. Panjonk, H.D. Steffens, B. Reznik, H. Buber, and H. Janett, Surface Analytical Examination of Welded Nickel Alloys by XPS, SNMS and SEM/EDX, Mikrochim. Acta, Vol 125 (No. 1–4), 1997, p 375–380 29. J. Monzo, J. Garcia-Politec, and J.L. Guinon, Influence of Elemental Sulfur and Mercaptans on Corrosion of Copper Strips in The ASTM D-130 Test by Means of Electron Microscopy and Energy Dispersive X- ray, J. Anal. Chem., Vol 341 (No. 10), 1991, p 606–610 30. I. Olefjord and A. Nylund, Surface Analysis of Oxidized Aluminum. 2. Oxidation Aluminum in Dry and Humid Atmosphere Studied By ESCA, SEM, SAM, and EDX, Surf. Interface Anal., Vol 21 (No. 5), 1994, p 290–297 31. N.H. Turner and J.A. Schreifels, Surface Analysis: X-Ray Photoelectron Spectroscopy and Auger Electron Spectroscopy, Anal. Chem., Vol 68, 1996, p 309R–331R 32. E.B. Varhegyi, I.V. Perczel, J. Gerblinger, M. Fleischer, H. Mexixner, and J. Giber, Auger and SIMS Study of Segregation and Corrosion Behavior of Some Semiconducting Oxide Gas-sensor Materials, Sens. Actuators B, Vol 19 (No. 1–3), 1994, p 569–572 33. D.B. Yang and D. Wolf, Interfacial Analysis of Copper Corrosion by Acrylic and Methycrylic Acids Using, XPS, Auger and Grazing angle FTIR, Surf. Interface Anal., Vol 23 (No. 5), 1995, p 276–288

34. S.J. Roosendaal, A.M. Vredenberg, and F.H.P.M. Habraken, The Influence of Adsorbed N in the Initial Stages of Oxidation of Iron, Surf. Sci., Vol 402–404 (No. 1–3), 1998, p 130–139 35. J.A. Schimpf, J.R. McBride, and M.P. Soriaga, Adsorbate-Catalyzed Layer-by-Layer Metal Dissolution in Halide-Free Solution, J. Phys. Chem., 1993, Vol 97, p 10,518–10,520 36. E. De Vito and P. Marcus, XPS Study of Passive Films Formed on Molybdenum-Implanted Austenitic Stainless Steel, Surf. Interface Anal., Vol 199 (No. 1–12), 1992, p 403–408 37. D. Thierry, D. Persson, C. Leygraf, N. Boucherit, and A.G. Hugot-le, Raman Spectroscopy and XPS Investigation of Anodic Films Formed on Fe-Mo Alloys In Alkaline Solutions, Corros. Sci., Vol 32 (No. 3), 1991, p 273–284 38. M.F. Fitzpatrick, J.S.G. Ling, and J.F. Watts, Failure Mechanism of Phosphated Adhesively Bonded Hot-dipped Galvanized Steel: A Small Area XPS Study, Surf. Interface Anal., Vol 29 (No. 2), 2000, p 131–138 39. C.R. Brundle, M. Grunze, U. Mader, and N. Blank, Detection and Characterization of Dimethylethanolamine-Based Corrosion Inhibitors at Steel Surfaces. I. The use of XPS and ToF-SIMS, Surf. Interface Anal., Vol 24 (No. 9), 1996, p 549–563 40. T. Fransen, R. Hofman, J.G.F. Westheim, I. Pouwel, and P.J. Gellings, FTIR and XPS Studies on Corrosion-resistant SiO2 Coatings As a Function of Humidity During Deposition, Surf. Interface Anal., Vol 24 (No. 1), 1996, p 1–6 41. J.L. Fang, Y. Li, X.R. Ye, Z.W. Wang, and Q. Liu, Passive Films and Corrosion Protection Due to Phosphoric Acid Inhibitors, Corrosion, Vol 49 (No. 4), 1993, p 266–271 42. V. Sirtori, F. Zambon, and L. Lombardi, XPS and Ellipsometric Characterization of Zinc-BTA Complex, J. Electron. Mater., Vol 29 (No. 4), 2000, p 463–467 43. F. Bentiss, M. Traisnel, L. Gengembre, and M. Lagreene, A New Triazol Derivative As Inhibitor of The Acid Corrosion of Mild Steel: Electrochemical Studies, Weight Loss Determination, SEM, XPS, Appl. Surf. Sci., Vol 152 (No. 3–4), 1999, p 237–249 44. P. Druska and H.H. Strehblow, Surface Analytical Examination of Passive Layers on Cu-Ni Alloys Part II. Acidic Solution, Corros. Sci., Vol 38 (No. 8), 1996, p 1369–1383 45. D. Schaepers and H.H. Strehblow, Surface Analytical Investigations of Electrochemically Formed Passive Layers on Binary Fe/ Al Alloys, J. Electrochem. Soc., Vol 142 (No. 7), 1995, p 2210–2218 46. P. Druska, H.H. Strehblow, and S. Golledge, A Surface Analytical Examination of Passive Layers on Cu/Ni Alloys .1. Alkaline Solutions, Corros. Sci., Vol 38 (No. 6), 1996, p 835–851 47. A. Rossi, C. Calinski, H.W. Hoppe, and H.H. Strehblow, A Combined ISS and XPS Investigation of Passive Film Formation On Fe53Ni, Surf. Interface Anal., Vol 18 (No. 4), 1992, p 269–276 48. R. Gasparac, C.R. Martin, E. StupnisekLisac, and Z. Mandic, In Situ and Ex Situ Studies of Imidazole and Its Derivatives As Copper Corrosion Inhibitors II. AC Impedance, XPS and SIMS Studies, J. Electrochem. Soc., Vol 147 (No. 3), 2000, p 991–998 49. H.J. Wouters, L.A. Butaya, and F.C. Adams, Applications of SIMS in Patina Studies on Bronze Age Copper Alloys, Anal. Chem., Vol 342 (No. 1–2), 1992, p 128–134

50. R. Singh, A.K. Tyagi, J.B. Gnanamoorthy, B. Raj, and S.K. Roy, Secondary Ion Mass Spectroscopy and Surface Prolifometric Characterization of Oxide Scales Developed Over Weld Metal, Heat Affected Zone, and Base Metal Regions of 9Cr-1Mo Steel Weldments, Mater. Sci. Technol., Vol 14 (No. 4), 1998, p 362–366 51. E. Cuyen, P. Lemberge, and P.E. Van, SIMS Depth Profiling of Cr-Conversion Layers on Aluminum, Electron Microscopy, General Interest Instrumentation, Vol 1, 1998, p 369–370 52. S. Yoshida and H. Ishida, An Investigation of The Thermal Stability of Undecylimidizole on Copper by FTIR Reflectance-Absorption Spectroscopy, Appl. Surf. Sci., Vol 89 (No. 1), 1995, p 39–47 53. M.L. McKelvy, T.R. Britt, B.R. Davis, J.K. Gillie, L.A. Lentz, A. Leugers, R.A. Nyguist, and C.L. Putzig, Infrared Spectroscopy, Anal. Chem., Vol 68, 1996, p 93R–160R 54. J. Fang, X.Ye, Q. Liu, Y. Li, and Z. Wang, XPS, AES, FTIR and Raman of An Anticorrosion Film on Steel, International Congress of Surface Finishing, Vol 2, 1992, p 1037–1045 55. G.L. Powel, A.G. Dobbins, S.S. Cristy, T.L. Clift, and H.M. Meyer, Study of The Oxidation of Uranium by External Diffuse Reflectance FTIR Spectroscopy Using Remote-Sensing and Evacuable Cell Techniques, Ninth International Conf. on FTIR Spectroscopy, Canada, 23–27 Aug 1993 56. A.V.R Kumar, R.K. Nigam, A.K. Srinavasta, and S.S. Monga, Mossbauer Spectroscopy and FTIR Spectroscopy of Corrosion Products, Obtained Beneath Mild Steel Coated with Different Phosphate Primers, Bull. Electrochem., Vol 15 (No. 3–4), 1999, p 160–163 57. J. Kasperek, D. Verchere, D. Jacquet, and N. Phillips, Analysis of the Corrosion Products on Galvanized Steels by FTIR Spectroscopy, Mater. Chem. Phys., Vol 56 (No. 3), 1998, p 205–213 58. A.K. Neufeld and I.S. Cole, Using FTIR Analysis to Detect Corrosion Products on the Surface of Metals Exposed to Atmospheric Conditions, Corrosion, Vol 53 (No. 10), 1997, p 788–799 59. S. Music, M. Gotic, and S. Popovic, X-Ray Diffraction and Fourier Transform Infrared Analysis of the Rust Formed by Corrosion of Steel in Aqueous Solutions, J. Mater. Sci., Vol 28 (No. 21), 1993, p 5744–5752 60. S. Shah, Characterization of Interphases in Corrosion By Ex Situ FTIR Spectroscopy, Corrosion 93, Vol 361, 1993, p 1–16 61. C.A. Melendres, Laser Raman Spectroscopy: Principles and Applications to Corrosion Studies, Electrochemical and Optical Techniques for the Study and Monitoring of Metallic Corrosion, Portugal, 9–21 July 1989 62. C.A. Melendres, Laser Raman and X-Ray Scattering Studies of Corrosion Films on Metals, 12th International Corrosion Congress, Houston, TX, 1993, p 19–24 63. J. Birnie, C. Craggs, D.J. Gardiner, and P.R. Graves, Ex Situ and In Situ Determination of Stress Distribution in Chromium Oxide Films by Raman Microscopy, Corros. Sci., Vol 33 (No. 1), 1992, p 1– 12 64. P.M.L Bonin, Electrochemical and Raman Spectroscopic Studies of The Influence of Chlorinated Solvents on The corrosion of Iron and Borate Buffer and In simulated Ground Water, Corrosion Science, 42 (11), 2000, p 1921–1939

65. R.K. Raman Singh, Laser Raman Spectroscopy: A Technique For Rapid Characterization of Oxide Scale Layers, Mater. Sci. Technol., Vol 14 (No. 5), 1998, p 373–376 66. J.E. Masler, In Situ Raman Spectroscopic Investigation of Aqueous Iron Corrosion at Elevated Temperatures and Pressures, J. Electrochem. Soc. Vol 147 (No. 7), 2000, p 2532–2542 67. J. Zhao, G. Frankel, and R.L. McCreery, Corrosion Protection of Untreated AA- 2024-T3 in Chloride Solutions by Chromate Conversion Coating Monitored with Raman Spectroscopy, J. Electrochem. Soc., Vol 145 (No. 7), 1998, p 2258–2264 68. D.L.A. deFria, S.V. Silva, and M.T. deOliveira, Raman Microscopy of Some Iron Oxides and Oxyhydroxides, J. Raman Spectrosc., Vol 28 (No. 11), 1997, p 873–878 69. M. Metikos-Hukovic, K. Furic, R. Babic, and A. Marinovic, Surface Enhanced Raman Scattering (SERS) of Benzotriazole Derivative Corrosion Inhibitor Prepared in Aqueous Media, Surf. Interface Anal., Vol 27 (No. 11), 1999, p 1016–1025 70. L.J. Oblonsky and T.M. Devine, Corrosion of Carbon Steels in CO2 Saturated Brine. A Surface Enhanced Raman Study, J. Electrochem. Soc., Vol 144 (No. 4), 1997, p 1252–1260 71. C.A. Malendres, Temperature Dependence of the Surface Enhanced Raman Spectroelectrochemistry of Iron in Aqueous Solutions, Electrochim. Acta, Vol 41 (No. 10), 1996, p 1727–1730 72. D.P. Schweinsberg, V. Otieno-Alego, G.A. Hope, and T. Notoya, An Electrochemical and SERS Study of the Effect of 1-[N,N-bis- (hydroxyethyl)aminomethyl]-benzotriole on the Acid Corrosion and Dezincification of 60/40 Brass, Corros. Sci., Vol 38 (No. 2), 1996, p 213–223 73. W.C. Baek, T. Kang, H.J. Sohn, and K.T. Young, In Situ Surface Enhanced Raman Spectroscopic Study on the Effect of Dissolved Oxygen on the Corrosion Film on Low Carbon Steel in 0.01 M NaCl Solution, Electrochim. Acta, Vol 46 (No. 15), 2001, p 2321–2325 74. H.H. Chan and M.J. Weaver, Vibrational Structural Analysis of Benzotriazol Adsorption and Phase Film Formation on Copper Using Surface Enhanced Raman Spectroscopy, Langmuir, Vol 15 (No. 9), 1999, p 3348–3355 75. J.L. Simpson and C.A. Melendres, Surface Enhanced Raman Spectroelectrochemical Studies of Corrosion Films on Iron in Aqueous Carbonate Solution, J. Electrochem. Soc., Vol 143 (No. 7), 1996, p 2146–2152 76. V. Shroeder and T.M. Devine, Surface Enhanced Raman Spectroscopic Study of Galvanostatic Reduction of the Passive Film on Iron, J. Electrochem. Soc., Vol 146 (No. 11), 1999, p 4061–4070 77. S.B. Torak, J. Labar, J. Injuk, and R.E.V. Grieken, X-Ray Spectroscopy, Anal. Chem., Vol 68, 1996, p 467R–485R 78. P. Schmuki, S. Virtanen, A.J. Davenport, and C.M. Vitus, Transpassive Dissolution of Cr and SputterDeposited Cr Oxides Studied by In Situ X-Ray Near Edge Spectroscopy, J. Electrochem. Soc., Vol 143 (No. 12), 1996, p 3997–4005 79. S. Pizzini, K.J. Roberts, I.S. Dring, R.J. Oldman, and D.C. Cupertino, Applications of X-Ray Absorption Spectroscopy to the Structural Characterization of Monodispersed Benzotriazole Coatings on Partly Oxidized Thin Films, J. Mater. Chem., Vol 3 (No. 8), 1993, p 811–819

80. J.S. Wainright, M.R. Antonio, and O.J. Murphy, The Oxidation State and Coordination Environment of Chromium in a Sealed Anodic Aluminum Oxide Film by X-Ray Absorption Spectroscopy, Corros. Sci., Vol 33 (No. 2), 1992, p 281–293 81. D.E.O.S. Carpenter, S.D. Carpenter, and J.P.G. Farr, Structure and Constitution of Zinc-Cobalt Electrodeposits, Corros. Sci., Vol 42 (No. 9), 2000, p 1599–1610 82. H.H. Strehblow and P. Borthen, EXAFS Analysis of Passive Films, Mater. Sci. Forum, Vol 185–188, 1995, p 301–311 83. G.D. Davis, W.C. Moshier, G.G. Long, D.R. Black, and B.A. Shaw, Passive Film Structure of Supersaturated Al-Mo Alloys, Conference: X-Ray Methods in Corrosion and Interface Electrochemistry, The Electrochemical Society, 1991, p 13–18 84. D. Lutzenkirchen-Hecht, C.U. Waligura, and H.H. Strehblow, An In Situ EXAFS Study of Corrosion Products, Corros. Sci., Vol 40 (No. 6), 1998, p 1037–1041 85. S.C. Chung, A.S. Lins, J.R. Chang, and H.C. Shih, EXAFS Study of Atmospheric Corrosion Products on Zinc at the Initial Stage, Corros. Sci., Vol 42, 2000, p 1599–1610 86. M. Balasubramanion and C.A. Melendres, Selective Site Occupancy Exhibited by Cr3+ and Cr6+ Incorporated into Electrochemically Deposited Nickel Hydroxide Films, Electrochem. Solid State Lett., Vol 2 (No. 4), 1999, p 167–169 87. G. Contini, A. Ciccioli, C. Laffon, P. Parent, and G. Polzonetti, NEXAFS Study of 2Mercaptobenzooxazole Adsorbed on Pt (111): Multilayer and Monolayer, Surf. Sci., Elsevier Science, Vol 412/413, 1998, p 158–165 88. A.I. Frenkel, and G.V. Korshin, EXAFS Studies of the Chemical State of Lead and Copper in Corrosion Products Formed on the Brass Surface in Potable Water, J. Synchrotron Radiation, Vol 6 (No. 3), 1999, p 653–655 89. D. Lutzenkirchen-Hecht and R. Frah, Time Resolved EXAFS Investigation of the Anodic Dissolution of Mo, J. Synchrotron Radiation, Vol 6 (No. 3), 1999, p 591–593 90. S. Suzuki, T. Suzuki, M. Kimura, Y. Takagi, K. Shinoda, K. Tohji, and Y. Woseda, EXAFS Characterization of Ferric Oxyhydroxide, Appli. Surf. Sci., Vol 169–170, 2001, p 109–112 91. G.G. Long, Surface X-Ray Absorption Spectroscopy, EXAFS and NEXAFS, for the in Situ and ex Situ Study of Electrodes, p 167–209, R. Varma and J.R. Selman, Ed. Techniques For Characterization of Electrodes And Electrochemical Processes, The Electrochemical Society, John Wiley & Sons, Inc., 1991 92. S. Shah, A. Ward, and A.J. Davenport, Characterization of Tellurium Based Conversion Coatings by Xray Absorption Near-Edge Structure and Fourier Transform Infrared Spectroscopy, Corrosion, Vol 51 (No. 1), 1995, p 67–70


References 1. S. Shah, Ed., Surface and Interface Characterization in Corrosion, NACE International, 1994 2. M. Riviere, Handbook of Surface and Interface Analysis Methods of Problem Solving, Marcel Dekker, Inc., 1998 3. R. Howland, L. Benatar, and C. Symanski, A Practical Guide to Scanning Probe Microscopy, DIANE Publishing Co., Collingdale, PA, 1998 4. J. Lipkowski and P.N. Ross, Imaging of Surfaces and Interfaces, John Wiley & Sons, Inc., 1999 5. A. Hubbard, The Handbook of Surface Imaging and Visualization, CRC Press, 1995 6. L.A. Bottomley, J.E. Coury, and P.N. First, Scanning Microscopy, Anal. Chem., Vol 68, 1996, p 185R– 230R 7. S. Shah, Corrosion: An Interfacial Process, Corrosion 91, “Paper 75,” 1991, p 1–22 8. T.G. Taranova, G.M. Grigorenko, V.A. Kostin, and V.V. Arsenyk, Applications of Scanning Electron Microscopy in Modern Material Science, Welding and Related Technologies, Welding International, Vol 14 (No. 9), 2000, p 745–748 9. Y. Ding and D.O. Northwood, Methods For Preparation of Cross-Sectional Scanning Electron Microscopic Specimens and Their Applications to Corroded Specimens of a Zirconium Alloy and Titanium Nitrogen Coated Stainless Steel, Mater. Charact., Vol 29 (No. 1), 1992, p 25–33 10. S.C. Tjong, Electron Microscope and Raman Characterization of The Surface Oxides Formed on The Fe-Cr Alloys at 400–800°C, Mater. Charact., Vol 26 (No. 1), 1991, p 29–44 11. R.K.R. Singh, A.K. Tyagi, K. Krishan, and J.B. Gnanamoorthy, Secondary Ion Mass Spectrometery and Scanning Electron Microscopic Characterization of Grain Boundary Oxide Ridges in 9Cr-1Mo Steels Having Different Silicon Contents, and Influence of Grain Size on Scale Spalling, Mater. Sci. Technol., Vol 10 (No. 7), 1994, p 592–598 12. L. Veleva and M. Luja, Scanning Electron Microscopic Characterization of Copper Corrosion Products, Br. Corros. J., Vol 34 (No. 1), 1999, p 34–36 13. S. Shah and M. Bernardo, Reusable Surgical Instruments: Metallurgy and Corrosion, Medical Product Sales, Vol 31 (No. 4), 2000, p 26–27 and Vol 31 (No. 6), 2000, p 26 14. P.J.R. Uwins, Environmental Scanning Electron Microscopy, Mater. Forum, Vol 18, 1994, p 51–57

15. K.P.J. Williams, I.C. Wilcock, I.P. Hayward, and A. Whitley, Applications of State-of- the-Art Raman Microscopy and Direct Raman Imaging. Progress in Corrosion, Mineralogy, and Materials Research, Spectroscopy, Vol 11 (No. 3), 1996, p 45–50 16. J. Kunze, V. Maurice, H.L. Klein, H.H. Strehblow, and P. Marcus, In Situ Scanning Tunneling Microscopic Study of The Anodic Oxidation of Copper (111) in 0.1 M NaOH, J. Phys. Chem. B, Vol 105 (No. 19), 2001, p 4263–4269 17. V. Maurice, H.H. Strehblow, and P. Maurice., In Situ STM Study of the Initial Stages of Oxidation of Cu (111) in Aqueous Solution, Surf. Sci., Vol 458 (No. 1–3), 2000, p 185–194 18. E.P. Friis, J.E.T. Anderson, L.L. Madsen, P. Moller, and J. Ulstrup, In Situ STM and AFM of The Copper Protein Pseudomonas aeruginosa azurin, J. Electroanal. Chem., Vol 431 (No. 1), 1997, p 35–38 19. D.J. Noll, P.G. Vanpatten, M.A. Nicholson, K. Booksh, and M.L. Myrick, Flow-Injection System for The Scanning Tunneling Microscope, Rev. Sci. Instrum., Vol 66 (No. 8), 1995, p 4150–4156 20. G.P. Bierwagen, R. Twite, G. Chen, and D.E. Tallman, Atomic Force Microscopy, Scanning Tunneling Microscopy and Electrochemical Characterization of Al Alloys, Conversion Coatings, and Primers Use for Aircraft, Prog. Org. Coatings, Vol 32 (No. 1–4), 1997, p 25–30 21. F.P. Zamborini, J.K. Campbell, and R.M. Crooks, Spectroscopic, Voltametric, and Electrochemical Scanning Tunneling Microscopic Study of Underpotentially Deposited Cu Corrosion and Passivation With Self-Assembled Organomercaptan Monolayers, Langmuir, Vol 14 (No. 3), 1998, p 640–647 22. C.J. Rubim, H. Jae, E. Henderson, and T.M. Cotton, Surface Enhanced Raman Scattering and Atomic Force Microscopy of Brass Electrodes in Sulfuric Acid Solution Containing Benzotriazole and Chloride Ion, Appl. Spectrosc., Vol 47 (No. 1), 1993, p 80–84 23. W. Kautek, M. Geu, M. Sabre, and P. Ahao, Multi-Method Analysis of Metal/Electrolyte Interface: Scanning Force Microscopy, Quartz Microbalance Measurements, FTIR and Grazing Incidence X-ray Diffractometry at Polycrystalline Electrode, Surf. Interface Anal., Vol 25 (No. 7–8), 1997, p 548–560 24. R.K.R. Singh and A.K. Tyagi, Secondary Ion Mass Spectrometry and Scanning Electron Microscopic Analysis of Grain Boundary Oxides in 9Cr-1Mo Steel and Influence of Grain Size on Scale Spalling, Mater. Sci. Technol., Vol 10 (No. 1), 1994, p 27–34 25. P. Lemoine and J.M. McLaughlin, Continuity and Topography of Ultrathin Diamond-Like Carbon Films Characterized by Scanning Electron Microscopy, Energy Dispersive X-ray Analysis and Atomic Force Microscopy, J. Vac. Sci. Technol., Vol 17 (No. 1), 1999, p 176–182 26. M. Bethencourt, F.J. Botana, J.J. Calvino, M. Marcos, J. Perez, and M. and A.R. Chacon, A SEM and EDS Insight Into The Corrosion Process of AA 5083 Al-Mg Alloy in NaCl Solutions, 14th Electron Microscopy Proc. of International Congress, B. Calderon, A. Hector, and M.Y. Jose, Ed. Vol 2 (No. 103–104), Institute of Physics Publishing, 1988 27. K. Forkel, D. Born, F.G. Wihsmann, and R. Stodolski, A SEM/EDXA Study on the Chemical Attack of Aqueous NaOH/KOH Solutions on the Surface of Metal-Metalloid Glass (Fe,Cr)80(P,C,Si)20 and Fe75Cr5P8C10SiO2, Respectively, Mikrochim. Acta, Vol 125 (No. 1–4), 1997, p 407–413 28. G. Panjonk, H.D. Steffens, B. Reznik, H. Buber, and H. Janett, Surface Analytical Examination of Welded Nickel Alloys by XPS, SNMS and SEM/EDX, Mikrochim. Acta, Vol 125 (No. 1–4), 1997, p 375–380

29. J. Monzo, J. Garcia-Politec, and J.L. Guinon, Influence of Elemental Sulfur and Mercaptans on Corrosion of Copper Strips in The ASTM D-130 Test by Means of Electron Microscopy and Energy Dispersive X- ray, J. Anal. Chem., Vol 341 (No. 10), 1991, p 606–610 30. I. Olefjord and A. Nylund, Surface Analysis of Oxidized Aluminum. 2. Oxidation Aluminum in Dry and Humid Atmosphere Studied By ESCA, SEM, SAM, and EDX, Surf. Interface Anal., Vol 21 (No. 5), 1994, p 290–297 31. N.H. Turner and J.A. Schreifels, Surface Analysis: X-Ray Photoelectron Spectroscopy and Auger Electron Spectroscopy, Anal. Chem., Vol 68, 1996, p 309R–331R 32. E.B. Varhegyi, I.V. Perczel, J. Gerblinger, M. Fleischer, H. Mexixner, and J. Giber, Auger and SIMS Study of Segregation and Corrosion Behavior of Some Semiconducting Oxide Gas-sensor Materials, Sens. Actuators B, Vol 19 (No. 1–3), 1994, p 569–572 33. D.B. Yang and D. Wolf, Interfacial Analysis of Copper Corrosion by Acrylic and Methycrylic Acids Using, XPS, Auger and Grazing angle FTIR, Surf. Interface Anal., Vol 23 (No. 5), 1995, p 276–288 34. S.J. Roosendaal, A.M. Vredenberg, and F.H.P.M. Habraken, The Influence of Adsorbed N in the Initial Stages of Oxidation of Iron, Surf. Sci., Vol 402–404 (No. 1–3), 1998, p 130–139 35. J.A. Schimpf, J.R. McBride, and M.P. Soriaga, Adsorbate-Catalyzed Layer-by-Layer Metal Dissolution in Halide-Free Solution, J. Phys. Chem., 1993, Vol 97, p 10,518–10,520 36. E. De Vito and P. Marcus, XPS Study of Passive Films Formed on Molybdenum-Implanted Austenitic Stainless Steel, Surf. Interface Anal., Vol 199 (No. 1–12), 1992, p 403–408 37. D. Thierry, D. Persson, C. Leygraf, N. Boucherit, and A.G. Hugot-le, Raman Spectroscopy and XPS Investigation of Anodic Films Formed on Fe-Mo Alloys In Alkaline Solutions, Corros. Sci., Vol 32 (No. 3), 1991, p 273–284 38. M.F. Fitzpatrick, J.S.G. Ling, and J.F. Watts, Failure Mechanism of Phosphated Adhesively Bonded Hot-dipped Galvanized Steel: A Small Area XPS Study, Surf. Interface Anal., Vol 29 (No. 2), 2000, p 131–138 39. C.R. Brundle, M. Grunze, U. Mader, and N. Blank, Detection and Characterization of Dimethylethanolamine-Based Corrosion Inhibitors at Steel Surfaces. I. The use of XPS and ToF-SIMS, Surf. Interface Anal., Vol 24 (No. 9), 1996, p 549–563 40. T. Fransen, R. Hofman, J.G.F. Westheim, I. Pouwel, and P.J. Gellings, FTIR and XPS Studies on Corrosion-resistant SiO2 Coatings As a Function of Humidity During Deposition, Surf. Interface Anal., Vol 24 (No. 1), 1996, p 1–6 41. J.L. Fang, Y. Li, X.R. Ye, Z.W. Wang, and Q. Liu, Passive Films and Corrosion Protection Due to Phosphoric Acid Inhibitors, Corrosion, Vol 49 (No. 4), 1993, p 266–271 42. V. Sirtori, F. Zambon, and L. Lombardi, XPS and Ellipsometric Characterization of Zinc-BTA Complex, J. Electron. Mater., Vol 29 (No. 4), 2000, p 463–467 43. F. Bentiss, M. Traisnel, L. Gengembre, and M. Lagreene, A New Triazol Derivative As Inhibitor of The Acid Corrosion of Mild Steel: Electrochemical Studies, Weight Loss Determination, SEM, XPS, Appl. Surf. Sci., Vol 152 (No. 3–4), 1999, p 237–249 44. P. Druska and H.H. Strehblow, Surface Analyti

S.D. Cramer, B.S. Covino, Jr., G.R. Holcomb and M. Ziomek-Moronz, Information Sources and Databases for the Corrosionist, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 999–1001

Information Sources Corrosionist

and

Databases

for

the

S.D. Cramer, B.S. Covino, Jr., G.R. Holcomb, M. Ziomek-Moroz; Albany Research Center, U.S. Department of Energy

Introduction REMARKABLE CHANGES in information technology have occurred since the 1987 edition of this Handbook was published. As a consequence, information and data that may not have been obtainable at that time are now routinely available by way of the World Wide Web. This article lists selected information sources and databases and the current (2003) way they are accessed. These lists were compiled from many other lists and from searches on the web. They represent the authors' choices of what are some of the more important corrosion sources on the web. They are not inclusive of all possible sites, which was not the authors' goal, but rather provide good starting points from which one can build. The information in this article has a fairly high “vapor pressure,” and its permanence can be fairly tenuous. On the other hand, the Internet addresses for major sites are not as likely to change since they provide continuity for the user. Information is provided in five areas: • • • • •

Societies and associations addressing corrosion and related topics Corrosion standards, specifications, recommended practices, and related topics Sources of corrosion information Corrosion databases and data compilations Other web resources

These areas frame much of the material on the web that addresses corrosion and related topics and provide portals to even more material.

S.D. Cramer, B.S. Covino, Jr., G.R. Holcomb and M. Ziomek-Moronz, Information Sources and Databases for the Corrosionist, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 999–1001 Information Sources and Databases for the Corrosionist S.D. Cramer, B.S. Covino, Jr., G.R. Holcomb, M. Ziomek-Moroz; Albany Research Center, U.S. Department of Energy

Societies and Associations Many of the leading societies and associations addressing corrosion and related topics and having a broad membership are listed in Table 1 with their web address. An effort has been made to provide a geographic distribution of organizations as well. The International Corrosion Council maintains perhaps the most comprehensive record of corrosion institutions in a list titled “Sources of Corrosion Information” that covers 40 countries. The list was last updated in 1996.

Table 1 Societies and associations addressing corrosion and related topics Organization ABRACO (Brazilian Corrosion Association) American Ceramic Society American Society for Microbiology AWWA (American Water Works Association) AISI (American Iron and Steel Institute) ASM International ASTM International Australasian Corrosion Association AWS (American Welding Society) China Anticorrosion Society Chinese Society for Corrosion and Protection Corrosion Institute of South Africa Corrosion Society of India Dechema (Society for Chemical Technology and Biotechnology) European Federation of Corrosion French Anticorrosion Centre Hungarian Corrosion Society Ibero-American Program of Science and Technology for Development Institute of Corrosion International Corrosion Council (ICC) ISE (International Society of Electrochemistry) Japanese Society of Corrosion Engineering Korean Institute of Metals and Materials Materials Research Society Mexican Society of Electrochemistry (SMEQ) NACE International National Paint and Coatings Association PRCI (Pipeline Research Council International) Polish Corrosion Society SSPC (Steel Structures Painting Council) Society for Biomaterials Society of Petroleum Engineers SPE (Society of Plastics Engineers) Spanish Council for Scientific Research (CSIC) Swedish Corrosion Institute ECS (The Electrochemical Society) The Institute of Materials, Minerals and Mining TMS (The Minerals, Metals and Materials Society) TAPPI (Trade Association of the Pulp and Paper Industry)

Web address www.abraco.org.br www.acers.org/ www.asmusa.org/ www.awwa.org/ www.steel.org/ www.asminternational.org/ www.astm.org/ www.corrprev.org.au/ www.aws.org/ www.china-anticorrosion.com/ www.cscp.org.cn/ www.corrosioninstitute.org.za/ www.naceindia.org/corrosion.html www.dechema.de/ www.efcweb.org/ www.cefracor.org/ www.chemres.hu/KKKI/hunkor/ahun.html www.cyted.org.ar/ www.icorr.demon.co.uk/ www.icc-net.org/ www.ise-online.org/ www.jcorr.or.jp/ www.kim.or.kr/ www.mrs.org/ www.sm-electroquimica.org www.nace.org/nace/index.asp www.paint.org/index.htm www.prci.com/ www.psk.org.pl/statut_a.htm www.sspc.org/ www.biomaterials.org/ www.spe.org/ www.4spe.org/ www.csic.es/english/principa.htm www.corr-institute.se/ www.electrochem.org/ www.iom3.org/index.htm www.tms.org/ www.tappi.org/


Corrosion Standards, Specifications, Recommended Practices, and Related Topics Standards, specifications, recommended practices, and test methods are essential to the organized progress of society and its goals. The standards organizations listed in Table 2 fall into three categories. The first have a mandate to provide standards across the breadth of society activities and are classified here as “broad spectrum.” Included in their activities are standards that address corrosion and related topics. The second are more focused organizations that address a particular constituency or interest group—such as TAPPI, which is concerned with the pulp and paper industry, or ACI International, which is focused on the concrete industry— and that more prominently address corrosion and its impact on the organizations' members and activities. The third are the organizations devoted specifically to all aspects of corrosion, such as NACE International. Table 2 Corrosion standards, specifications, recommended practices, and related topics Source ACI International American Association of State and Highway Transportation Officials (AASHTO) ANSI (American National Standards Institute) American Water Works Association (AWWA) American Welding Society (AWS) ASTM International (American Society for Testing and Materials) British Standards Institution (BSI) Canadian General Standards Board (CGSB) Department of Defense Single Stock Point (DODSSP) European Committee for Standardization (CEN) International Standards Organization (ISO) NACE International National Institute of Standards Technology (NIST) Polish Committee for Standards

and

SSPC (Steel Structures Painting Council) TAPPI (Trade Association of the Pulp and Paper Industry)

Nature of standards Concrete, including corrosion Transportation, including corrosion Broad spectrum

Web address www.concrete.org/ www.transportation.org/

Water, including corrosion

www.awwa.org/

Welding, including thermal spray Broad spectrum, including corrosion Broad spectrum, including corrosion Broad spectrum, including corrosion U.S. Military standards and specification Broad spectrum, including corrosion Broad spectrum, including corrosion Corrosion Measurement, documentary, conformity, and information Broad spectrum, including corrosion Paints, coatings, and linings Pulp and paper, including corrosion

www.aws.org/ www.astm.org/

www.ansi.org/

www.bsi-global.com/ www.pwgsc.gc.ca/cgsb/ www.dodssp.daps.mil www.cenorm.be/ www.iso.ch/ www.nace.org/ www.nist.gov/ www.pkn.pl/ www.sspc.org/ www.tappi.org/


Sources of Corrosion Information The web provides many sources of information on corrosion, ranging from corrosion fundamentals, the forms of corrosion, and how corrosion is detected and measured, to the multitude of ways equipment and structures are protected from corrosion damage, and how society and the environment are protected from the consequences of corrosion damage. Selected sites are given in Table 3, representing a reasonably broad range of interests and providing portals to other sources of information. This list is, as they say, “only the tip of the iceberg,” and the authors appreciate the fact that many readers have their own list of favorites. Table 3 Sources of corrosion information Source ACI International American Galvanizers Association ASM International Copper Development Association Corrosion Doctors CorrosionSource Cost of Corrosion CSIRO (Commonwealth Scientific and Industrial Research Organization) EPRI (Electric Power Research Institute) Gas Institute International Corrosion Council International Lead and Zinc Research Organization (ILZRO) LaQue Center for Corrosion Technology McMaster University, Walter W. Smeltzer Corrosion Laboratory Metal Finishing Industry MIT, Department of Materials Science and Engineering, H.H. Uhlig Corrosion Laboratory Montana State University, Center for Biofilm Engineering Nickel Development Institute NIST (National Institutes of Science and Technology) Ohio State University, Fontana Corrosion Center Royal Institute of Technology, Materials Science and Engineering, Division of Corrosion Science

Web address www.aci-int.org/ www.galvanizeit.org/ www.asminternational.org/ www.copper.org/ www.corrosion-doctors.org/ www.corrosionsource.com/ www.corrosioncost.com/ www.csiro.au/ www.epri.com/ www.gri.org/ www.icc-net.org/ www.ilzro.org/ www.laque.com/ www.mse.eng.mcmaster.ca/resource/corrlab.htm www.finishing.com/ www.dmse.mit.edu/UhligLab/

www.erc.montana.edu/ www.nidi.org/ www.nist.gov/ www.er6.eng.ohio-state.edu/~frankel/fcc/ www.met.kth.se/corr/

TWI (The Welding Institute) UMIST Corrosion and Protection Centre University of Delaware, Marine BiologyBiochemistry University of Louisiana at Lafayette, Corrosion Research Center University of Virginia, Center for Electrochemical Science and Engineering Transportation Research Information Services (TRIS), Transportation Research Board, National Academy of Sciences Office of Scientific and Technical Information (OSTI), U.S. Department of Energy National Physical Laboratory, High Temperature Corrosion and Coatings Sci.chem.electrochem Newsgroup

www.twi.co.uk/ www.cp.umist.ac.uk/CPC/ www.ocean.udel.edu/cms/mbbfac.html www.engr.louisiana.edu/crc/ www.virginia.edu/cese/ www4.trb.org/trb/tris.nsf/

www.osti.gov/energycitations/

www.npl.co.uk/npl/cmmt/htc/ www.groups.google.com/groups?group=sci.chem..electrochem/


Corrosion Databases and Data Compilations Few situations in science and engineering exist where all the data needed are immediately available from direct experience. Corrosion— with the possible combinations of multiple materials, the multitude of corrosion agents in a variety of strengths over a range of temperatures, and other environmental factors—presents a daunting array of scenarios. Databases and data compilations represent the collective knowledge gleaned from a multitude of individual experiences. Databases and data compilations, supported by an understanding of fundamentals, move one toward goals with efficiency and economy in less-than-perfect situations; that is, most situations. The databases and data compilations listed in Table 4 provide fundamental information on a wide range of materials, their properties, their use and handling, and for evaluating their application. The data compilations are books that may or may not be available in electronic form. Their publishers are listed under contact information. Table 4 Selected corrosion databases and data compilations Publisher or source Databases American Chemical Society ASM International

Title or description

Contact information

CAS (Chemical Abstracts Services) substance databases Alloy Center (including Alloy Finder, Data Sheets & Diagrams, Corrosion Data, Coatings Data, and Materials Property Data) Phase Diagrams (binary and ternary

www.info.cas.org/substance.html www.asminternational.org/matinfo

www.asminternational.org/bookstore

Cambridge Scientific Abstracts (CSA) Elsevier Science

ESM Software

NACE International National Institute of Standards and Technology (NIST)

University of Akron U.S. Bureau of Labor Statistics Various

ASSET Project (U.S. Department of Energy, Office of Industrial Technologies) Data compilations ASM International

Marcel Dekker

NACE International

phase diagrams on CD-ROM) Corrosion and materials science bibliographic databases Dechema Corrosion Handbook: Corrosive Agents and Their Interaction with Materials TAPP (thermochemical and physical properties of pure chemical compounds) ConSur (Corrosion Data Survey Database) Access to over 80 NIST scientific and technical databases, including NIST molten salts (properties of 320 inorganic single salts and 4000 mixtures) Hazardous chemical database Consumer price indexes for the United States beginning as early as 1913 Material Safety Data Sheets

Alloy selection system for elevated temperatures; predicts metal loss in gaseous atmospheres from 250– 1150 °C (480–2100 °F) Atlas of Stress-Corrosion and Corrosion Fatigue Curves, 1990 Handbook of Corrosion Data, 2nd ed., 1995 Corrosion Resistance Tables: Metals, Nonmetals, Coatings, Mortars, Plastics, Elastomers, and Linings and Fabrics, 4th ed., 1995 Atlas of Electrochemical Equilibria in Aqueous Solutions, 1974 Compilation and Use of Corrosion Data for Alloys in Various HighTemperature Gases, 1999 Corrosion Data Survey: Metals Section, 6th ed. 1985 Corrosion Data Survey: Nonmetals Section, 5th ed., 1975

www.csa.com www.elsevier.com/

www.esm-software.com/tapp/

www.nace.org www.nist.gov/srd/

www.ull.chemistry.uakron.edu/erd/ www.bls.gov/cpi/

Available from various sources. A directory of Internet sources can be found at www.ilpi.com/msds www.oit.doe.gov/chemicals/

www.asminternational.org/

www.dekker.com/

www.nace.org/


Other Web Resources There are some web sites that simply need to be visited to view the links that they provide. That is the purpose of Table 5. These are sites that provide connections to a wealth of information relevant to corrosion and its related topics. Table 5 Other web resources Site name Case Western Reserve University Corrosion Doctors CorrosionSource Steel-related technologies

UMIST Corrosion Information Server University of Delaware Library Sci.chem.electrochem Newsgroup

Description Electrochemical science and technology information resource Corrosion links, including simulation software Useful links (in corrosion) Electrochemistry, corrosion science, and microbiologically influenced corrosion resources Internet resources in corrosion and materials Internet resources in materials science Newsgroup community focused on electrochemistry

Web address www.electrochem.cwru.edu/estir/inet.htm www.corrosion-doctors.org/Links.htm www.corrosionsource.com/links.htm www.steelynx.net

www2.umist.ac.uk/corrosion/cis/links.htm www2.lib.udel.edu/subj/masc/internet.htm See Table 3

S.D. Cramer, B.S. Covino, Jr., G.R. Holcomb and M. Ziomek-Moronz, Conventions and Definitions in Corrosion and Oxidation, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 1002–1008

Conventions and Definitions in Corrosion and Oxidation Stephen D. Cramer, Bernard S. Covino, Jr., Gordon R. Holcomb, and Małgorzata Ziomek-Moroz, Albany Research Center, U.S. Department of Energy; Jack Tinnea, Tinnea Associates

Introduction COMMON CONVENTIONS AND DEFINITIONS in corrosion, cathodic protection, electricity, and oxidation that can result in confusion and some degree of aggravation are presented in this article. The definitions and

conventions have been grouped but, of course, are not exclusive to any one category. A more comprehensive set of definitions is provided in the “Glossary of Terms” in this Volume.

S.D. Cramer, B.S. Covino, Jr., G.R. Holcomb and M. Ziomek-Moronz, Conventions and Definitions in Corrosion and Oxidation, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 1002–1008 Conventions and Definitions in Corrosion and Oxidation Stephen D. Cramer, Bernard S. Covino, Jr., Gordon R. Holcomb, and Małgorzata Ziomek-Moroz, Albany Research Center, U.S. Department of Energy; Jack Tinnea, Tinnea Associates

Corrosion and Electrochemical Cells Anion. An anion is an ion that has a negative charge. An anion will move toward the anode in an electric field. Anode. In an electrochemical cell, the anode is the electrode where oxidation takes place. See also the section “Cell Types and Conditions for Commercial and Industrial Processes” in this article. Cathode. In an electrochemical cell, the cathode is the electrode where reduction takes place. See also the section Cell Types and Conditions for Commercial and Industrial Processes in this article. Cation. A cation is an ion that has a positive charge. A cation will move toward the cathode in an electric field. Corrosion. In its broadest sense, corrosion is the deterioration of a metallic or nonmetallic material by reaction with its environment involving processes other than strictly mechanical or thermal processes. Applied to metallic materials, corrosion is the deterioration of a material by chemical or electrochemical reactions between the material and its environment. Corrosion Cell. A corrosion cell is typically a galvanic cell that results from a lack of homogeneity in a material or its environment. There are cases where a corrosion cell can be an electrolytic cell, for example, in stray current corrosion. Corrosion Potential. The corrosion potential is the mixed electrode potential of a freely corroding material relative to a reference electrode, also called the open-circuit potential. Current: Sign Convention. Anodic currents and current densities are positive, and cathodic currents and current densities are negative. This corresponds to the convention that current (referred to here as conventional current) is the flow of positive charge; electrons move in a direction that is counter to the flow of current. When displaying anodic and cathodic current or current density values on logarithmic coordinates, absolute values of the current or current density are used and the cathodic values are clearly indicated. The direction of conventional current in an external circuit is defined by the cell type: Cell type Direction of conventional current Reversible No net current From the cathode (+) to the anode (-) Galvanic Electrolytic From the cathode (-) to the anode (+) From the cathode (+) to the anode (-) Corrosion Current and Current Density: Symbol Conventions. I is commonly used to represent current, and i is commonly used to represent current density or flux. Some authors use J or j to represent current density. Subscripts may be added to these symbols to further distinguish them. For example a, c, or o may be used for anodic, cathodic, and exchange current, respectively. The unit of I is the ampere (A), which is a base unit in the SI system or some multiple. Units of i are ampere per unit area, such as A/ m2 (A/ft2). Dissociation. is the breakdown of dissolved compounds in a solvent into their charged constituents or components. These charged constituents or ions are the substances that carry charge through the solution. As a consequence of dissociation, the properties of acids, bases, and salts dissolved in water or other polar solvents are very different from those of the undissolved materials.

To explain dissociation, Arrhenius theory states that an acid is a substance that produces hydrogen ions (H+), and a base is a substance that produces hydroxide ions (OH-) in water. Consequently, an acid solution contains an excess of hydrogen ions and a base an excess of hydroxide ions. The reactions that produce these ions are: acid: HA → H+ + A-

(Eq 1)

base: catOH → cat+ + OH-

(Eq 2)

salt: catA → cat+ + A-

(Eq 3) +

+

where A is an anion and cat is a cation different from H . In contrast, Brønsted-Lowry proton transfer theory states that an acid is a substance that dissociates to produce a proton and a base is a substance that accepts a proton: acid → proton + base

(Eq 4)

HCl → H+ + Cl-

(Eq 5)

A proton (H+) does not exist by itself in solution. Protons combine with water to form the hydrated hydrogen ion (H3O+), are known as the hydronium ion. Thus, Eq 5 becomes: HCl + H2O → H3O+ + Cl-

(Eq 6)

Arrhenius theory applies to aqueous solutions. The Brønsted-Lowry proton transfer theory applies to all media. It is the more general and important theory, particularly when considering the chemistry of substances in solutions other than water. An example in aqueous solution is the salt ammonium chloride (NH4Cl). Aqueous solutions of NH4Cl are acidic. The Brønsted-Lowry explanation shows that the ammonium ion ( proton donor and:

) is a

(Eq 7) According to Brønsted-Lowry theory, an acid such as H2O, HCl, H2S, HCN, and HNO3 as well as cations and anions such as H3O+, , HS-, , and can be electrically neutral. Bases do not need to contain OH and can be H2O, NH3, NH2OH, CH3NH2, N2H4, OH-, Cl-, CN-, CH3COO-, or [Zn(H2O)3OH]+. Electrochemical Cell. It is convenient to consider an electrochemical cell as consisting of two independent electrode or half-cell reactions. The anode supports an anodic reaction, and the cathode supports a cathodic reaction. Electrons carry charge through a metallic circuit (often an external circuit), and ions carry charge through the electrolyte. Cell Types and Conditions for Commercial and Industrial Processes. Table 1 lists conditions that exist in different cell types and the commercial and industrial processes that correspond to these cell types. See the article “Introduction to Corrosion for Constructive Purposes” in this Volume for more information. Table 1 Cell conditions for commercial and industrial electrode processes Cell type

Cell potential Reversible ecell,rev ecell < Galvanic ecell,rev Electrolytic ecell > ecell,rev

Net current I=0 I ≠ 0 (spontaneous) I ≠ 0 (impressed current or applied potential)

Process Polarity Anode Cathode + … + Batteries, fuel cells +

-

Electrolytic polishing Electroplating Electrochemical machining

Corrosion

ecell ~ 0

Icorr (spontaneous)

-

+

Electrochemical refining Corrosion Chemical-mechanical

planarization Currents and Potentials at Electrodes in Different Cell Types. The currents and potentials at two reversible electrodes, copper and silver, are given in Table 2 as an example of the conditions that exist for different cell types. The electrode systems chosen are a copper electrode immersed in a copper sulfate solution at unit activity of Cu2+ and a silver electrode immersed in a silver nitrate solution at unit activity of Ag+. This choice of electrodes (or half-cells) avoids the complexity that arises with many electrodes of commercial interest where oxides or other stable corrosion products form on the surface. It also avoids the complexity that arises with reactions involving chemical species from different redox systems, such as zinc in an aqueous solution where the reduction reaction can be the evolution of hydrogen or the reduction of dissolved oxygen. The solutions are connected by a porous barrier and are at a temperature of 25 °C (77 °F). See the article “Introduction to Corrosion for Constructive Purposes” in this Volume for more information. Table 2 Electrical conditions at each electrode Cell potential

Reversible

Electrode reactions Cu Ag Ia = Ic = Io Ia = Ic = Io

Galvanic

Ia > Ic

Ia < Ic

ecell < ecell,rev

E>

E
Ic

ecell > ecell,rev

E

Corrosion

Ia > Ic

Ia < Ic

ecell = 0

Cell type

Electrode potentials, ESHE Cu2+/Cu Ag+/Ag

ecell,rev = 0.459 V

= 0.340

= 0.799

E> E< Ia, current associated with the anodic reaction at the electrode; Ic, current associated with the cathodic reaction at the electrode; Io, exchange current. E, the potential of the electrode with reference to the standard hydrogen electrode (SHE); ecell, the potential between external connections to the cell Electrode Potential: Temperature Coefficient Sign Convention. The thermal temperature coefficient is positive when an increase in temperature produces an increase in electrode potential; that is, the electrode potential shifts in the positive or more noble direction. Electrode Potential: Measuring An Unknown Potential. In terms of measuring the potential of an electrode in an electrolyte, connect the reference electrode to the ground (negative) terminal of the measuring device (electrometer, multimeter, etc.) and the electrode of interest to the high (positive) terminal. A negative potential in this configuration means the unknown electrode potential is negative to the reference electrode and vice versa. The cell notation for this configuration (see Electrode Reactions and Cell Notation in this article) is: (Eq 8) A simple check of measuring device polarity is achieved with any battery. Connect the measuring device lead used for the reference electrode to the negative battery terminal. Connect the measuring device lead used for the electrode of interest to the positive battery terminal. The measuring device should show a positive potential. If it does not, the polarity switch of the measuring device should be moved to the positive position. Electrode Reactions and Cell Notation. An electrochemical cell is composed of at least two half-cell reactions. An anodic reaction is: A → An+ + ne-

(Eq 9)

A cathodic reaction is: Cm+ + me- → C

(Eq 10)

to produce the net or overall cell reaction: mA + nCm+ → mAn+ + nC

(Eq 11)

The corresponding half-cell reactions are described by the following notation: anodic reaction: A|An+(aq, aAn+) cathodic reaction:

(Eq 12a)

Cm+(aq, aCm+)|C

(Eq 12b)

These half-cell reactions are combined as follows to denote the net cell reaction: A|An+(aq, aAn+)||Cm+(aq, aCm+)|C

(Eq 13)

In this notation, the anode or oxidizing reaction is always on the left, and the cathode or reducing reaction is always on the right. Variations of this notation allow for a metallic electrode that does not react electrochemically, for dissolved gases, and for solid phases in equilibrium with the solution. The two vertical lines represent a cell where the liquid junction potential between two aqueous solutions (aq) has been eliminated or is assumed to be small. Liquid junction potentials are often small by comparison to the cell potential. The cell potential for the net cell reaction (Eq 14), that is, the difference between the half-cell potentials for the cathode (right) and anode (left) reactions, is: ecell = Eright - Eleft

(Eq 14)

where the potential on the right is half-cell (Eq 12bb) and the potential on the left is half-cell (Eq 12aa), and both are expressed as reduction potentials. The electrochemical cell notation representing the measurement of the unknown potential of an electrode of interest with respect to the standard hydrogen electrode (SHE) is given by: (Eq 15a) SHE||electrode reaction of interest

(Eq 15b)

where f = 1 means the fugacity is one and a = 1 means unity activity. Electrode Potential: Sign Convention. The electropositive direction of electrode potential implies an increasing oxidizing condition at the electrode. The electropositive direction is also known as the noble direction because the corrosion potentials of noble metals such as gold and platinum are more electropositive than those of more active base metals such as magnesium, zinc, and iron. The electronegative direction is associated with an increasing reducing condition at the electrode. The electronegative direction is also known as the active direction, because the corrosion potentials of active metals such as magnesium and zinc are substantially more electronegative than those of noble metals. Half-cell electrode reactions written as reduction potentials follow this convention, adopted in 1953 by the International Union of Pure and Applied Chemistry, and yield positive potentials for more oxidizing conditions and negative potentials for more reducing conditions (Ref 1). Electrolyte. An electrolyte is a nonmetallic electrical conductor where the current is carried by ions instead of electrons. Exchange Current. The exchange current (Io) is the common value of the anodic and cathodic currents when an electrode is at equilibrium with its environment Io = Ia = -Ic

(Eq 16)

Half-Cell Reactions. A half-cell reaction is the electrochemical reaction that describes charge transfer at an electrode and involves either a reduction or oxidation reaction. Ion. An ion is an atomic or molecular particle having a net electric charge. Reference Electrode Potentials: Values And Conversion Factors. Table 3 lists reduction electrode potentials for some common reference electrodes and the conversion factors to change measured potential values from that of the measuring electrode to the SHE (standard hydrogen electrode) and SCE (saturated calomel electrode) scales.

Table 3 Reference electrode potentials and conversion factors Electrode

Potential at 25 °C (77 °F)

+

V vs. SHE 0.000

Thermal temperature coefficient, mV/°C

Add the following value in volts to convert from measurements with the given reference electrode(a): To SHE scale To SCE scale 0.000 -0.241

+0.87 (Pt)/H2(f = 1)/H (a = 1), SHE +0.235 +0.25 +0.235 -0.006 Ag/AgCl/1 M KCl +0.250 … +0.250 +0.009 Ag/AgCl/0.6 M Cl (seawater) +0.22 +0.288 +0.047 Ag/AgCl/0.1 M Cl- +0.288 +0.22 +0.241 0.000 Hg/Hg2Cl2/sat KCl, +0.241 SCE +0.280 +0.59 +0.280 +0.039 Hg/Hg2Cl2/1 M KCl +0.334 +0.79 +0.334 +0.093 Hg/Hg2Cl2/0.1 M KCl +0.300 +0.90 +0.30 +0.06 Cu/CuSO4sat +0.616 … +0.616 +0.375 Hg/HgSO4 (a) An electrode potential of 1.000 V vs. SCE would be 1.000 + 0.241 = 1.241 V vs. SHE. An electrode potential of -1.000 V vs. SCE would be -1.000 + 0.241 = -0.759 V vs. SHE. Source: Ref 2 Overpotential is the current-producing potential difference between a nonequilibrium electrode potential and its corresponding equilibrium value for a definite electrode reaction, η = E - Eo (Ref 3, Fig. 1). Overvoltage and overpotential are often used interchangeably (Ref 4).

Fig. 1 Evans diagram showing the relationship between electrode potential and overpotential Polarization is defined similarly to the overpotential; that is, the difference between the potentials of an electrode with and without current, E - Eo (Ref 3, 4). The common forms of electrode polarization are activation polarization (ηact); concentration polarization (ηconc); and resistance polarization (ηr). Polarization is the sum of these contributions: η = ηact + ηconc + ηr

(Eq 17)

A term overvoltage (ε) has been defined similarly to overpotential to describe polarization (the current-voltage relationship) around the mixed potential (Ecorr) for corrosion reactions (Ref 5); that is, ε = E - Ecorr. The important distinction is that overpotential is related to the reversible potential of an electrode reaction, and overvoltage is related to the mixed potential of a corrosion reaction. Some of the confusion regarding polarization occurs because the term overvoltage can be used interchangeably with overpotential (Ref 4) or in the context of “overvoltage” (Ref 5).

Polarization Data: Conventions for Displaying. On rectangular coordinates, positive values are plotted above the origin on the y-axis (ordinate axis) and to the right of the origin on the x-axis (abscissa axis). On logarithmic coordinates, increasing values are plotted from left to right on the x-axis and from bottom to top on the y-axis. Current or current density is plotted on the x- axis, and potential is plotted on the y-axis. Current or current density is plotted using either rectangular coordinates or logarithmic coordinates, depending on the range of current or current density values to be displayed. Absolute values of current or current density are used with logarithmic coordinates. Typically, current or current density values are plotted as amperes (A) or amperes per square meter (A/m2). Potential is plotted using rectangular coordinates. Typically, potentials are converted to the SHE scale or the SCE scale and are plotted as volts (V) measured on this scale.

References cited in this section 1. K.B. Oldham, and G. Mansfeld, Corrosion, CORRA, Vol 27, 1971, p 434 2. “Conventions Applicable to Electrochemical Measurements in Corrosion Testing,” G- 3, Annual Book of ASTM Standards, 1989, p 46–53 3. D.A. Jones, Principles and Prevention of Corrosion, Macmillan, New York, 1992, p 79–85 4. G. Wranglén, An Introduction to Corrosion and Protection of Metals, Chapman and Hall, London, 1985, p 10 5. D.A. Jones, Principles and Prevention of Corrosion, Macmillan, New York, 1992, p 145–146


Cathodic Protection Cathodic Protection (CP) is the reduction of corrosion by polarizing a structure to a sufficiently cathodic potential so that corrosion in the most anodic area(s) on the structure is/are reduced to an acceptably low level. Impressed-current CP (ICCP) protects the structure by making the structure the cathode (negative electrode) in an electrolytic cell consisting of an external source of direct current (dc) power and either a consumable or a nonconsumable anode (positive electrode) of material such as high-silicon cast iron, graphite, mixed metal oxide, or catalyzed titanium (Fig. 2a).

Fig. 2 Simplified layouts for cathodic protection systems. (a) Impressed current cathodic protection system (b) Galvanic or sacrificial cathodic protection system Sacrificial CP (SACP)—also known as galvanic CP—protects the structure by coupling the structure to a consumable anode of a more active material such as zinc or magnesium (Fig. 2a). An ICCP system is an electrolytic cell. A sacrificial CP system is a galvanic cell. Both systems polarize the structure to be protected in an electronegative direction. A properly designed CP system causes all areas of the protected structure to be made cathodic to the most anodic site on the structure before CP. Cathodic Protection Testing Protocol. The working convention of many CP practitioners deviates from standard electrochemical convention. Two areas where this occurs are in current flow and meter configuration. “Conventional current” flows from positive to negative. This is opposite the flow of electrons. If one uses the flow of electrons as current flow and standard electrochemical convention in voltmeter connections for potential readings, some of the confusion in CP practice can be avoided. Voltmeters used by early pipeline CP workers were analog and only read upscale. Potential surveyors for buried steel structures would almost always encounter readings that were negative with respect to a CSE (saturated copper-sulfate reference electrode). Given that the meters of the day read only up the scale, workers were in the habit of connecting the reference electrode to the positive terminal of the voltmeter and the structure to the negative and then adding the “-” in their heads. Also, by holding to the idea that “negative” always goes to the structure, it was thought that this practice helped to avoid problems with connecting ICCP systems backwards.

In time, the “upscale” meters were provided polarity switches, which were then followed by the autoranging/autopolarity digital meters of today. However, the old practice of connecting the CSE to the positive terminal of the voltmeter and the structure to the negative and changing the polarity by notation persists as of this printing in some sectors. The practice of connecting the reference to the positive voltmeter terminal can lead to confusion in other areas. For example, SACP anodes are selected because they are more electronegative than the structure to be protected. The structure thus is positive with respect to the SACP anodes. In contrast, the rectifier positive terminal in ICCP is connected to the anodes while the negative terminal is connected to the structure. In both cases, however, the flow of conventional current in the external circuit is away from the structure or, equivalently, the flow of electrons is towards the structure (Fig. 2). It would be easy to think that, since the galvanic SACP anodes are electronegative to the structure while the ICCP anodes are electropositive, the current flow in the two CP systems is reversed. This is not true. CP Protection Criteria. Early in the application of CP to pipeline protection, polished steel rods were used as “most anodic” references. The logic behind this approach was that if the pipe were polarized to the potential of the polished steel (zero volts between the pipe and rod), it would then be polarized to the level of most anodic sites. Reference electrodes, especially CSE, eventually displaced polished steel rods as the reference of choice. Likewise, standards were developed by organizations that defined protection criteria not only for buried pipelines but also for metals in other environments ranging from those submerged in seawater to those embedded in concrete. Four common criteria are used for unalloyed ferrous materials (Ref 6, 7): • • • •

The IR-free (“instant off”) potential of buried materials must be more negative than -850 mV versus CSE. The “on potential” of buried materials is 300 mV more negative than the free corrosion potential in soil, similar to criterion 1. IR-free polarization of buried materials must be 100 mV more electronegative than the corrosion potential measured either by current interrupt or depolarization techniques. The protection current is sufficient to maintain the material at a condition that meets any of the above criteria in periodic testing.

CP Design: Area Basis for Current Densities. Current densities are employed in rule-of- thumb approaches to the design of cathodic protection systems. An assumption is made about the current density required to protect the structure. The selected density value (mA/m2) is multiplied by the nominal surface area (m2) of the structure in contact with the electrolyte to provide an estimate of the total required current. When a metal surface is new, there is a fairly good relationship between nominal surface area and the metal surface area in contact with the surrounding electrolyte. With age, the metal may corrode, roughness develops, and the actual metal surface area in contact with the electrolyte may increase to several times the “nominal” surface. In a similar fashion, when a metal to be protected is covered with a protective coating, the initial surface area in contact with the electrolyte is much less than the nominal surface area. With time, the coating will degrade and the surface area of the metal in contact with the electrolyte will increase. Current densities also are important in the operation of CP anodes. In reinforced concrete, it is recommended (Ref 8) that the maximum sustained current density on the anode be less than 108 mA/m2. In practice, some structure owners (Ref 9) operate their anodes at current densities as low as 2 mA/m2. Evans Diagram: Impressed Current CP (ICCP) in Neutral or Basic Environment. Figure 3(a) shows a plot of potential versus log current (Evans diagram) for an ICCP system in a neutral or basic environment. Only activation and concentration polarization are included in the structure corrosion kinetics. The structure anode reaction is the dissolution of iron; the structure cathode reaction is the reduction of oxygen. The intersection between the structure anodic polarization curve and the structure cathodic polarization curve determines the unprotected structure corrosion potential (Ecorr,S) and the unprotected structure corrosion current (Icorr,S).

Fig. 3 Evans diagrams for cathodic protection (CP) systems. (a) Diagram for an impressed current CP system protecting a steel structure in an oxygenated neutral or basic environment using an inert anode. (b) Diagram for a galvanic or sacrificial CP system protecting a steel structure in an oxygenated neutral or basic environment using a consumable zinc anode. (c) Diagram for a galvanic or sacrificial CP system protecting a steel structure in an acidic environment using a consumable zinc anode.

Coupling the structure to a nonconsumable inert anode powered by a rectifier shifts the structure to a more electronegative potential (Ecp,S). This shift reduces the structure anodic current to Icp,S, the corrosion current with cathodic protection. Because the cathodic reaction at the structure is diffusion limited and thus highly polarized, the current required by this shift in potential changes very little from Icorr,S. The reaction at the inert anode is the oxidation of hydroxyl ions to form oxygen. As power is applied to the structure, the potential of the inert anode shifts in the electropositive direction to EA, the potential of the inert anode at a current equal to Icorr,S. The ΔE shown, the IR drop in the solution, is a function of the partial current collecting on the structure (Ion,S) (equivalent to Icorr,S), and the electrolyte resistivity (R), in the vicinity of the reference electrode used to evaluate the system. Because of this IR drop, the structure appears to be polarized to Eon,S. The rectifier output is given by V = EA - Eon,S. When the current is interrupted, an IR free measurement of the structure potential is achieved after inductive artifacts have dissipated (typically ~0.5 seconds). This potential is the “instant off” potential and equivalent to Ecp,S. With the current off, the structure potential decays back toward Ecorr,S. The value of this decay or depolarization over a period of time, typically 4 to 24 h, is a measure of how well the structure is protected by the application of CP. The reduction in the corrosion current is given by ΔI = Icorr,S - Icp,S. In some situations, say deep well configurations, the environment surrounding the anodes may become acidic and reactions other than those shown may dominate. These reactions include gas generation, which if not vented, may greatly increase the electrical resistance at the anode. Evans Diagram: Galvanic or Sacrificial Cathodic Protection (SACP) in Neutral or Basic Environment. Figure 3(b) shows an Evans diagram for SACP in a neutral or basic environment. The structure part of this diagram and the notation is the same as that used for the ICCP example. Only the anode has been changed, from an inert anode for ICCP to a consumable zinc anode for SACP. Again, only activation and concentration polarization are included in the structure corrosion kinetics. The structure anode reaction is the dissolution of iron; the structure cathode reaction is the reduction of oxygen. The intersection between the structure anodic polarization curve and the structure cathodic polarization curve determines the unprotected structure corrosion potential (Ecorr,S) and the unprotected structure corrosion current (Icorr,S). Coupling the structure to a consumable zinc anode shifts the structure to a more electronegative potential (Ecp,S). This shift reduces the structure anodic current to Icp,S, the corrosion current with cathodic protection. Because the cathodic reaction at the structure is diffusion limited and highly polarized, the current required by this shift in potential changes very little from Icorr,S. The reaction at the sacrificial anode is the oxidation of zinc to form zinc ions in solution. The potential of the zinc anode shifts in the electropositive direction to Ecp,S, the potential of the zinc anode at a current equal to Icorr,S. Just as in the case of ICCP, there is typically an IR potential drop between the structure and the zinc anode, and the structure and the zinc anode are at somewhat different potentials. However, for clarity, this has not been shown in Fig. 3(b). As in the case of ICCP, when the current is interrupted, an IR free value of the structure potential, which is the “instant off potential” and equivalent to Ecp,S, is measured. With the current off, the structure potential decays back towards Ecorr,S. The value of this decay or depolarization over a period of time, typically 4 to 24 h, is a measure of how well the structure is protected by the application of CP. The reduction in the corrosion current is given by ΔI = Icorr,S - Icp,S. Of course, these determinations cannot be made on SACP systems where the consumable anode and structure cannot be isolated from each other. Evans Diagram: Galvanic or Sacrificial Cathodic Protection (SACP) in Acidic Environment. Figure 3(c) is an Evans diagram for SACP in an acidic environment using a consumable zinc anode. Note that both the zinc anode and the structure are polarized by the process as indicated by the dashed lines showing the combined reduction and oxidation rates. For clarity, similar details were not included in Fig. 3(a) and 3(b). Only activation polarization is included in the structure corrosion kinetics. The structure anode reaction is the dissolution of iron; the structure cathode reaction is the reduction of hydrogen ions. The intersection between the structure anodic polarization curve and the structure cathodic polarization curve determines the unprotected structure corrosion potential (Ecorr(Fe)) and the unprotected structure corrosion current (Icorr(Fe)). Coupling the structure to a consumable zinc anode shifts the structure to a more electronegative potential (Ecorr(couple)). This shift reduces the structure anodic current to Icp(Fe), the corrosion current with cathodic protection. Because the cathodic reaction at the structure is not diffusion limited, the current required by this shift in potentials increases substantially to Icorr(couple), representing the increase in the cathodic current due to

hydrogen reduction on the zinc anode and the increase in the anodic current due to dissolution of the zinc anode. The reaction at the consumable zinc anode is the oxidation zinc to form zinc ions in solution. The potential of the freely corroding zinc electrode in the acidic environment is Ecorr(Zn) and the corrosion current for the zinc at this potential is Icorr(Zn) . When the zinc anode is coupled with the structure, the potential shifts in the electropositive direction to Ecorr(couple). Just as in the case of ICCP, there is typically an IR potential drop between the structure and the zinc anode, and the structure and the zinc anode are at somewhat different potentials. However, for clarity, this has not been shown in Fig. 3(c). As in the case of ICCP, when the current is interrupted, an IR free value of the structure potential is measured, which is the “instant off potential” and equivalent to Ecorr(couple). With the current off, the structure potential decays back towards Ecorr(Fe). The value of this decay or depolarization over a period of time, typically 4 to 24 h, is a measure of how well the structure is protected by the application of CP. The reduction in the corrosion current is given by ΔI = Icorr(Fe) - Icp(Fe). Of course, these determinations cannot be made on SACP systems where the consumable anode and structure cannot be isolated from each other.

References cited in this section 6. W. von Baeckmann, W. Schwenk, and W. Prinz, Handbook of Cathodic Protection, 3rd ed., Gulf Publishing Co, 1997, p 103–105 7. “Control of External Corrosion on Underground or Submerged Metallic Piping Systems,” Standard Recommended Practice No. RP0169-2002, NACE International, Houston, TX, 2002 8. “Cathodic Protection of Reinforcing Steel in Atmospherically Exposed Concrete Structures,” RP029090, NACE International, Houston, TX, 1990, 11 pages 9. B.S. Covino, Jr., S.D. Cramer, S.J. Bullard, G.R. Holcomb, J.H. Russell, W.K. Collins, H.M. Laylor, and C.B. Cryer, “Performance of Metalized Zinc Anodes for Cathodic Protection of Reinforced Concrete Bridges— Final Report,” FHWA-OR-RD-02-10, Federal Highway Administration, U.S. Department of Transportation, Washington D.C., 2002. Available at http:// www.odot.state.or.us/tddresearch/reports/ zn%20anode%20.pdf and at www.osti.gov/ servlets/purl/804079-JhD8ED/native/


Electricity Classifying Materials by Conductivity: Conductors (Metals), Semiconductors, and Insulators. Materials can be considered conductors (metals), semiconductors, and insulators. Metals are good conductors of electrons. An electric current flows under the influence of an electric or changing magnetic field in a metal. In insulators, there is no electric current flow under an electric field. In terms of the energy-band theory (Ref 10), insulators have completely full valence bands and cannot carry electrical current. Metals have partially filled valence bands and thus can carry current. In materials where the energy gap between the valence band and the band

immediately above it is small (less than about 2 eV at room temperature), electrons can be thermally excited from the valence band to the next band. This creates partially filled bands, and, thus, electrical current can flow. Such materials are semiconductors. Electronic defects in nonstoichiometric semiconductors can be negative (electrons) or positive (holes) and are termed n-type and p-type semiconductors, respectively. Conductivity. Conduction of electrical charge through an electrolyte results from the mobility of ions. Through a metal, conduction results from the mobility of electrons, and through a gas, conduction results from the mobility of gaseous ions and electrons. Conductivity, also called specific conductance, is a measure of these transport processes and a property of a homogeneous material. It is the inverse of resistivity and has the units of (Ω · length)-1 or S/ m. The siemens (S) is the conductance of a material such that one ampere current will flow with an applied potential of one volt; S = 1/Ω Electrochemical impedance (Z) is a complex number defined as: Z(jω) = Z′(jω) + jZ″(jω)

(Eq 18)

where Z′ is the real or in-phase component of impedance; Z″ is the imaginary or out-of-phase component of impedance; j2 = -1, and ω is the frequency in radian/s and = 2πf where f is the frequency in Hertz (Hz). The impedance magnitude is |Z|2 = (Z′)2 + (Z″)2. The phase angle between the real and imaginary components of impedance is: θ = arctan (Z″/Z′). The admittance Y = 1/Z = Y′ + jY″. Impedance can be modeled in a linear constant parameter system as: Z = R + jX

(Eq 19)

where R is resistance; X is reactance X = ωL - 1/ωC; and L is inductance and C is capacitance. Resistivity. Resistivity, also called specific resistance (ρ), is a property of a homogeneous material. It is computed from the resistance (R) of a volume of the material having unit length (l) and unit cross-sectional area (A). Its value is ρ = RA/l, and it has the units of Ω · l, such as Ω · cm.

Reference cited in this section 10. M.A. Omar, Elementary Solid State Physics, Addison-Wesley, Reading, MA, 1975, p 210–212


Oxidation Crystal Structure. A crystal structure is the arrangement of atoms or ions in a regularly repeating pattern, such that the atoms or ions are located either on the points of a Bravais lattice or in a fixed relation to those points (Ref 11). A Bravais lattice is a set of points in space such that each point has identical surroundings. There are 14 possible arrangements or crystal structures (Table 4 and Fig. 4). The name of the crystal structure is a combination of the Bravais lattice and system; that is, body-centered cubic.

Table 4 Crystal structures System

Bravais lattices

Shape Axial lengths a=b=c

Cubic

Simple

Angles α = β = γ = 90°

Tetragonal

Face-centered Simple

a=b≠c

α = β = γ = 90°

Orthorhombic

Body-centered Simple

a≠b≠c

α = β = γ = 90°

Face-centered Simple Simple Simple

a=b=c a=b≠c a≠b≠c

α = β = γ ≠ 90° α = β = 90°; γ = 120° α = β = 90°; γ ≠ 90°

Base-centered Simple

a≠b≠c

α ≠ β ≠ γ ≠ 90°

Body-centered

Body-centered Base-centered

Rhombohedral (or trigonal) Hexagonal Monoclinic

Triclinic Refer to Fig. 4. Source: Ref 11

Fig. 4 Crystal axes, interaxial angles, and unit-cell edge length Defect Structures: Kröger-Vink Notation. A notation used to describe point defects in inorganic compounds was developed by Kröger and Vink (Ref 12, 13, 14). In this system, the type of defect is indicated by a major symbol, and the site it occupies is indicated by a subscript. A superscript indicates the charge relative to a normal lattice, where ′ designates an effective negative charge, and · designates an effective positive charge. An x indicates a zero effective charge. Some examples are:

Designation Meaning Doubly charged oxygen vacancy Cr (+3) on a Ni(+2) site Oxygen on an oxygen site (a normal anion) Metal on a metal site (a normal cation) Metal interstitial Mi Metal on a metal site with an extra electron Free electron (concentration of electrons n) e′ · Electron hole (concentration of holes p) h Defect Structures. Common point defect structures found in inorganic compounds are briefly described (Ref 15). Kröger-Vink notation is used to describe the defect equilibria. Schottky defects are found in stoichiometric compounds and consist of equivalent concentrations of anion and cation defects. An example would be pairs of oxygen and metal vacancies: (Eq 20) Frenkel defects are found in stoichiometric compounds and are formed when a metal ion on a regular cation site is moved to an interstitial site and leaves behind a vacancy. An example for singly charged interstitials and vacancies is: (Eq 21) Defects associated with oxygen-deficient or metal excess nonstoichiometric compounds can include extra oxygen vacancies or extra metal interstitials. An example of each is: (Eq 22) (Eq 23) Defects associated with oxygen excess or metal-deficient nonstoichiometric compounds can include extra oxygen interstitials or extra metal vacancies. An example of each is: (Eq 24) (Eq 25) Dopants can create a wide range of defects where the foreign atom takes the place of the metal cation in the normal lattice or as an interstitial. Intrinsic ionization of electrons from the valence band to the conduction band create e′ and h· pairs. This typically occurs in stoichiometric compounds. Parabolic Rate Equations. The oxidation of many metals at high temperatures can be described with parabolic kinetics. Often Wagner's oxidation theory is invoked to explain the behavior. It is worthwhile to note the assumptions used in Wagner's theory so that its limitations can be recognized when describing more complex systems (Ref 16, 17, 18, 19): • • • •

• • •

Lattice diffusion of the reacting ions or atoms, or the transport of electrons through the dense compact scale, are rate determining. Lattice diffusion occurs through the migration of point defects. These defects migrate independently from each other. Reactions at the phase boundaries are rapid, and thermodynamic equilibria are established at the metal/oxide and oxide/oxygen interfaces. The driving force for the oxidation reaction is the free energy change that occurs on forming the oxide from the metal and gaseous oxygen. Thus a gradient in oxygen activity exists across the scale, with the activity at the metal/oxide interface determined from the metal/oxide equilibrium. Deviations from stoichiometry are small. O2 solubility is negligible. Scales are thick enough to neglect charged surfaces.

The major deviations from the Wagner theory include easy diffusion paths, the formation of voids and porosity, and the effects of stresses in the scale. Parabolic Rate Constant Evaluation: Use Δm2 versus t (not Δm versus t0.5). The kinetics of high-temperature corrosion experiments are often reported in terms of a parabolic rate law and a parabolic rate constant, kp. One then has a choice to determine kp from the slope of the experimental change in mass (Δm) and time (t) data plotted as either Δm versus t0.5 or Δm2 versus t. For the simplest form of the parabolic rate expression, Δm2 = kpt, the choice is immaterial. However, for most metallic materials, there is an initial transient period of fast kinetics that is best described by (Δm - Δmi)2 = kp(t - ti). Pieraggi (Ref 20) has shown that systematic errors in determining kp arise when there is this initial transient period and the slope of Δm versus t0.5 is used instead of the slope of Δm2 versus t.

References cited in this section 11. B.D. Cullity, Elements of X-Ray Diffraction, 2nd ed., Addison-Wesley Publishing Company, Reading, MA, 1978, p 34–35 12. F.A. Kröger and H.J. Vink, Solid State Phys., Vol 3, 1956, p 307 13. F.A. Kröger, The Chemistry of Imperfect Crystals, North-Holland, Amsterdam, 2nd ed., 1974 14. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, New York, 1988, p 34–36 15. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, New York, 1988, p 43–51 16. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, New York, 1988, p 163–164 17. C. Wagner, Z. Phys. Chem., Vol B21, 1933, p 25 18. C. Wagner, Atom Movements, American Society Metals, 1951, p 153 19. C. Wagner, Prog. Solid State Chem., Vol 10, 1975, p 3 20. B. Pieraggi, Oxidation of Metals, Vol 27 (No. 3/4), 1987, p 177–185


References 1. K.B. Oldham, and G. Mansfeld, Corrosion, CORRA, Vol 27, 1971, p 434 2. “Conventions Applicable to Electrochemical Measurements in Corrosion Testing,” G- 3, Annual Book of ASTM Standards, 1989, p 46–53

3. D.A. Jones, Principles and Prevention of Corrosion, Macmillan, New York, 1992, p 79–85 4. G. Wranglén, An Introduction to Corrosion and Protection of Metals, Chapman and Hall, London, 1985, p 10 5. D.A. Jones, Principles and Prevention of Corrosion, Macmillan, New York, 1992, p 145–146 6. W. von Baeckmann, W. Schwenk, and W. Prinz, Handbook of Cathodic Protection, 3rd ed., Gulf Publishing Co, 1997, p 103–105 7. “Control of External Corrosion on Underground or Submerged Metallic Piping Systems,” Standard Recommended Practice No. RP0169-2002, NACE International, Houston, TX, 2002 8. “Cathodic Protection of Reinforcing Steel in Atmospherically Exposed Concrete Structures,” RP029090, NACE International, Houston, TX, 1990, 11 pages 9. B.S. Covino, Jr., S.D. Cramer, S.J. Bullard, G.R. Holcomb, J.H. Russell, W.K. Collins, H.M. Laylor, and C.B. Cryer, “Performance of Metalized Zinc Anodes for Cathodic Protection of Reinforced Concrete Bridges— Final Report,” FHWA-OR-RD-02-10, Federal Highway Administration, U.S. Department of Transportation, Washington D.C., 2002. Available at http:// www.odot.state.or.us/tddresearch/reports/ zn%20anode%20.pdf and at www.osti.gov/ servlets/purl/804079-JhD8ED/native/ 10. M.A. Omar, Elementary Solid State Physics, Addison-Wesley, Reading, MA, 1975, p 210–212 11. B.D. Cullity, Elements of X-Ray Diffraction, 2nd ed., Addison-Wesley Publishing Company, Reading, MA, 1978, p 34–35 12. F.A. Kröger and H.J. Vink, Solid State Phys., Vol 3, 1956, p 307 13. F.A. Kröger, The Chemistry of Imperfect Crystals, North-Holland, Amsterdam, 2nd ed., 1974 14. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, New York, 1988, p 34–36 15. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, New York, 1988, p 43–51 16. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, New York, 1988, p 163–164 17. C. Wagner, Z. Phys. Chem., Vol B21, 1933, p 25 18. C. Wagner, Atom Movements, American Society Metals, 1951, p 153 19. C. Wagner, Prog. Solid State Chem., Vol 10, 1975, p 3 20. B. Pieraggi, Oxidation of Metals, Vol 27 (No. 3/4), 1987, p 177–185

Corrosion Rate Conversion,, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 1028

Corrosion Rate Conversion Table 1 Relationships among some of the units commonly used for corrosion rates d is metal density in grams per cubic centimeter (g/cm3) Factor for conversion to Unit mdd g/m2/d μm/yr mm/yr mils/yr in./yr 0.1 36.5/d 0.0365/d 1.144/d 0.00144/d Milligrams per square decimeter per day 1 (mdd) 10 1 365/d 0.365/d 14.4/d 0.0144/d Grams per square meter per day (g/m2/d 0.0274d 0.00274d 1 0.001 0.0394 0.0000394 Microns per year (μm/yr) 27.4d 2.74d 1000 1 39.4 0.0394 Millimeters per year (mm/yr) 0.696d 0.0696d 25.4 0.0254 1 0.001 Mils per year (mils/yr) 696d 69.6d 25,400 25.4 1000 1 Inches per year (in./yr) Source: G. Wranglén, An Introduction to Corrosion and Protection of Metals, Chapman and Hall, 1985, p. 238

Glossary of Terms A absorption. A process in which fluid molecules are taken up by a liquid or solid and distributed throughout the body of that liquid or solid. Compare with adsorption. accelerated corrosion test. Method designed to approximate, in a short time, the deteriorating effect under normal long-term service conditions. acicular ferrite. A highly substructured nonequiaxed ferrite formed upon continuous cooling by a mixed diffusion and shear mode of transformation that begins at a temperature slightly higher than the transformation temperature range for upper bainite. It is distinguished from bainite in that it has a limited amount of carbon available; thus, there is only a small amount of carbide present. acid. A chemical substance that yields hydrogen ions (H+) when dissolved in water; also a substance that dissociates to produce a proton (H+) in any medium, that is, a proton donor. Compare with base. acid embrittlement. A form of hydrogen embrittlement that may be induced in some metals by acid. acid rain. Atmospheric precipitation with a pH below 5.6 to 5.7. Burning of fossil fuels for heat and power is the major factor in the generation of oxides of nitrogen and sulfur, which are converted into nitric and sulfuric acids in atmospheric moisture (fogs and precipitation). See also atmospheric corrosion. acrylic. Resin polymerized from acrylic acid, methacrylic acid, esters of these acids, or acrylonitrile. activation. The changing of a passive surface of a metal to a chemically active state. Contrast with passivation.

active. The negative direction of electrode potential. Also used to describe corrosion and its associated potential range when the electrode potential is more negative than adjacent potentials where the corrosion rate is depressed; that is, the passive region. active metal. A metal ready to corrode, or readily corroded. active potential. The potential of a corroding material. activity. A measure of the chemical potential of a substance, where the chemical potential is not equal to concentration, that allows mathematical relations equivalent to those for ideal systems to be used to correlate changes in experimentally measured quantities with changes in chemical potential. activity (ion). The ion concentration corrected for deviations from ideal behavior. Concentration multiplied by activity coefficient. activity coefficient. A characteristic of a quantity expressing the deviation of a solution from ideal thermodynamic behavior; often used in connection with electrolytes. addition agent. A substance added to a solution for the purpose of altering or controlling a process. Examples include wetting agents in acid pickling, brighteners or antipitting agents in plating solutions, and inhibitors. adsorption. The surface retention of solid, liquid, or gas molecules, atoms, or ions by a solid or liquid. Compare with absorption. aeration. (1) Exposing to the action of air. (2) Causing air to bubble through. (3) Introducing air into a solution by spraying, stirring, or a similar method. (4) Supplying or infusing with air, as in sand or soil. aeration cell (oxygen cell). See differential aeration cell. age hardening. Hardening by aging, usually after rapid cooling or cold working. aging. A change in the properties of certain metals and alloys that occurs at ambient or moderately elevated temperatures after hot working or a heat treatment (quench aging in ferrous alloys, natural or artificial aging in ferrous and nonferrous alloys) or after a cold- working operation (strain aging). The change in properties is often, but not always, due to a phase change (precipitation), but never involves a change in chemical composition of the metal or alloy. See also age hardening, artificial aging, natural aging, overaging, precipitation hardening, precipitation heat treatment, quench aging, and strain aging. alclad. Composite wrought product comprising an aluminum alloy core having on one or both surfaces a metallurgically bonded aluminum or aluminum alloy coating that is anodic to the core and thus electrochemically protects the core against corrosion. alkali metal. A metal in group IA of the periodic system—namely, lithium, sodium, potassium, rubidium, cesium, and francium. They form strongly alkaline hydroxides, hence the name. alkaline. (1) Having properties of an alkali. (2) Having a pH greater than 7. alkaline cleaner. A material blended from alkali hydroxides and such alkaline salts as borates, carbonates, phosphates, or silicates. The cleaning action may be enhanced by the addition of surface-active agents and special solvents. alkyd. Resin used in coatings. Reaction products of polyhydric alcohols and polybasic acids. alkylation.

(1) A chemical process in which an alkyl radical is introduced into an organic compound by substitution or addition. (2) A refinery process for chemically combining isoparaffin with olefin hydrocarbons. alligatoring. (1) Pronounced wide cracking over the entire surface of a coating having the appearance of alligator hide. (2) The longitudinal splitting of flat slabs in plane parallel to the rolled surface. Also called fishmouthing. alloy plating. The codeposition of two or more metallic elements. alpha ferrite. See ferrite. alpha iron. The body-centered cubic form of pure iron, stable below 910 °C (1670 °F). alternate-immersion test. A corrosion test in which the specimens are intermittently exposed to a liquid medium at definite time intervals. aluminizing. Forming of an aluminum or aluminum alloy coating on a metal by hot dipping, hot spraying, or diffusion. amalgam. An alloy of mercury with one or more other metals. ammeter. An instrument for measuring the magnitude of electric current flow. amorphous solid. A rigid material whose structure lacks crystalline periodicity; that is, the pattern of its constituent atoms or molecules does not repeat periodically in three dimensions. See also metallic glass. amphoteric. A term applied to oxides and hydroxides that can act basic toward strong acids and acidic toward strong alkalis. Substances that can dissociate electrolytically to produce hydrogen or hydroxyl ions according to conditions. anaerobic. Free of air or uncombined oxygen. anchorite. A zinc-iron phosphate coating for iron and steel. anion. A negatively charged ion that migrates through the electrolyte toward the anode under the influence of a potential gradient. See also cation and ion. annealing. A generic term denoting a treatment, consisting of heating to and holding at a suitable temperature, followed by cooling at a suitable rate, used primarily to soften metallic materials, but also to simultaneously produce desired changes in other properties or in microstructure. The purpose of such changes may be, but is not confined to, improvement of machinability, facilitation of cold work, improvement of mechanical or electrical properties, and/or increase in stability of dimensions. When the term is used by itself, full annealing is implied. When applied only for the relief of stress, the process is properly called stress relieving or stress-relief annealing. anode. The electrode of an electrolyte cell at which oxidation occurs. Electrons flow away from the anode in the external circuit. It is usually at this electrode that corrosion occurs and metal ions enter solution. Contrast with cathode. anode corrosion. The dissolution of a metal acting as an anode. anode corrosion efficiency. The ratio of the actual corrosion (weight loss) of an anode to the theoretical corrosion (weight loss) calculated by Faraday's law from the quantity of electricity that has been generated by the anode during the corrosion.

anode effect. The effect produced by polarization of the anode in electrolysis. It is characterized by a sudden increase in voltage and corresponding decrease in amperage due to the anode becoming virtually separated from the electrolyte by a gas film. anode efficiency. The current efficiency at the anode. anode film. (1) The solution in immediate contact with the anode, especially if the concentration gradient is steep. (2) The outer layer of the anode itself. anode polarization. See polarization. anodic cleaning. Electrolytic cleaning in which the work is the anode. Also called reverse- current cleaning. anodic coating. A film or scale on a metal surface resulting from an electrolytic treatment at the anode. anodic inhibitor. A chemical substance or mixture that prevents or reduces the rate of the anodic or oxidation reaction. See also inhibitor. anodic polarization. The change of the electrode potential in the noble (positive) direction due to current flow. See also polarization. anodic protection. Imposing an external electrical potential to polarize a metal into the passive region where dissolution rates are low (applicable only to metals that show active- passive behavior). Contrast with cathodic protection. anodic reaction. Electrode reaction equivalent to a transfer of positive charge from an electronic to an ionic conductor. An anodic reaction is an oxidation process. A common example in corrosion is: Me → Men+ + ne-. anodizing. Forming an anodic coating or conversion coating on a metal surface by anodic oxidation; frequently applied to aluminum. anolyte. The electrolyte adjacent to the anode in an electrolytic cell. antifouling. Intended to prevent fouling of underwater structures, such as the bottoms of ships. antipitting agent. An addition agent for electroplating solutions to prevent the formation of pits or large pores in the electrodeposit. aqueous. Pertaining to water; an aqueous solution is made by using water as a solvent. artificial aging. Aging above room temperature. See also aging. Compare with natural aging. atmospheric corrosion. The gradual degradation or alteration of a material by contact with substances present in the atmosphere, such as oxygen, moisture, acidic gases, acid precipitation, basic particulated, carbon dioxide, sulfur, and chlorine compounds. sulfur and chlorine compounds. atomic weight. A number assigned to each chemical element that specifies the average mass of its atoms. Because an element may consist of two or more isotopes having different masses, the atomic weight is the average of the masses of the naturally occurring isotopes proportional to their occurrence. The mean weight of the atom of an element in relation to 12C = 12.000. austenite. A solid solution of one or more elements in face-centered cubic iron. Unless otherwise designated (such as nickel austenite), the solute is generally assumed to be carbon.

austenitizing. Forming austenite by heating a ferrous alloy into the transformation range (partial austenitizing) or above the transformation range (complete austenitizing). When used without qualification, the term implies complete austenitizing. auxiliary anode. In electroplating, a supplementary anode positioned so as to raise the current density on a certain area of the cathode and thus obtain better distribution of plating. auxiliary electrode. An electrode commonly used in polarization studies to pass current to or from a test electrode. It is usually made from a noncorroding material; often called the counterelectrode. B backfill. Material placed in a drilled hole to fill space around anodes, vent pipe, and buried components of a cathodic protection system. bainite. A metastable aggregate of ferrite and cementite resulting from the transformation of austenite at temperatures below the pearlite range but above Ms, the martensite start temperature. Bainite formed in the upper part of the bainite transformation range has a feathery appearance; bainite formed in the lower part of the range has an acicular appearance resembling that of tempered martensite. banded structure. A segregated structure consisting of alternating nearly parallel bands of different composition, typically aligned in the direction of primary hot working. base. A chemical substance that yields hydroxyl ions (OH-) when dissolved in water; also, a proton receptor in any medium. Compare with acid. base metal. (1) The metal present in the largest proportion in an alloy; for example, brass is a copper-base alloy. (2) An active metal that readily oxidizes, or that dissolves to form ions. (3) The metal to be brazed, cut, soldered, or welded. (4) After welding, that part of the metal that was not melted. beach marks. Macroscopic progression marks on a fatigue fracture or stress-corrosion cracking surface that indicate successive positions of the advancing crack front. The classic appearance is of irregular elliptical or semielliptical rings, radiating outward from one or more origins. Beach marks (also known as clamshell marks or arrest marks) are typically found on service fractures where the part is loaded randomly, intermittently, or with periodic variations in mean stress or alternating stress. See also striation. biaxial stress. See principal stress (normal). biological corrosion. Deterioration of metals as a result of the metabolic activity of microorganisms. See also microbiologically influenced corrosion. bipolar electrode. An electrode in an electrolytic cell that is not mechanically connected to the power supply, but is so placed in the electrolyte, between the anode and cathode, that the part nearer the anode becomes cathodic and the part nearer the cathode becomes anodic. Also called intermediate electrode. bituminous coating. Coal tar or asphalt-base coating. black liquor. The liquid material remaining from pulpwood cooking in the soda or sulfate papermaking process. black oxide. A black finish on a metal produced by immersing it in hot oxidizing salts or salt solutions. blister. A raised area, often dome shaped, resulting from (1) loss of adhesion between a coating or deposit and the base metal or (2) delamination under the pressure of expanding gas trapped in a metal in a nearsubsurface zone. Very small blisters may be called pinhead blisters or pepper blisters.

blow down. (1) Injection of air or water under high pressure through a tube to the anode area for the purpose of purging the annular space and possibly correcting high resistance caused by gas blocking. (2) In connection with boilers or cooling towers, the process of discharging a significant portion of the aqueous solution in order to remove accumulated salts, deposits, and other impurities. blue brittleness. Brittleness exhibited by some steels after being heated to a temperature within the range of about 200 to 370 °C (400 to 700 °F), particularly if the steel is worked at the elevated temperature. blushing. Whitening and loss of gloss of a usually organic coating caused by moisture. Also called blooming. brackish water. (1) Water having salinity values ranging from approximately 0.5 to 17 parts per thousand. (2) Water having less salt than seawater, but undrinkable. breakdown potential (Eb) The least noble potential where pitting or crevice corrosion, or both, will initiate and propagate. See also critical pitting potential. brightener. An agent or combination of agents added to an electroplating bath to produce a smooth, lustrous deposit. brine. Typically, water containing a higher concentration of dissolved salt than that of ordinary ocean water. brittle fracture. Separation of a solid accompanied by little or no macroscopic plastic deformation. Typically occurs by rapid crack propagation with less expenditure of energy than for ductile fracture. burning. (1) Permanently damaging a metal or alloy by heating to cause either incipient melting or intergranular oxidation. See also overheating. (2) In grinding, getting the work hot enough to cause discoloration or to change the microstructure by tempering or hardening. C calcareous coating or deposit. A layer consisting of a mixture of calcium carbonate and magnesium hydroxide deposited on surfaces being cathodically protected because of the increased pH adjacent to the protected surface. calomel electrode. An electrode widely used as a reference electrode of known potential in electrometric measurement of acidity and alkalinity, corrosion studies, voltammetry, and measurement of the potentials of other electrodes. See also electrode potential, reference electrode, and saturated calomel electrode. calorizing. Imparting resistance to oxidation to an iron or steel surface by heating in aluminum powder at 800 to 1000 °C (1470 to 1830 °F). carbonitriding. A case hardening process in which a suitable ferrous material is heated above the lower transformation temperature in a gaseous atmosphere of such composition as to cause simultaneous absorption of carbon and nitrogen by the surface and, by diffusion, create a concentration gradient. The process is completed by cooling at a rate that produces the desired properties in the workpiece. carburizing. Absorption and diffusion of carbon into solid ferrous alloys by heating, to a temperature usually above Ac3, in contact with a suitable carbonaceous material. A form of case hardening that produces a carbon gradient extending inward from the surface, enabling the surface layer to be hardened either by quenching directly from the carburizing temperature or by cooling to room temperature, then reaustenitizing and quenching. case hardening. A generic term covering several processes applicable to steel that change the chemical composition of the surface layer by absorption of carbon, nitrogen, or a mixture of the two and, by diffusion, create a concentration gradient. The outer portion, or case, is made substantially harder than the inner portion, or

core. The processes commonly used are carburizing and quench hardening; cyaniding; nitriding; and carbonitriding. The use of the applicable specific process name is preferred. CASS test. See copper-accelerated salt-spray test. cathode. The electrode of an electrolytic cell at which reduction occurs. Electrons flow toward the cathode in the external circuit. Typical cathodic processes are cations taking up electrons and being discharged, oxygen being reduced, and the reduction of an element or group of elements from a higher to a lower valence state. Contrast with anode. cathode efficiency. Current efficiency at the cathode. cathode film. The solution in immediate contact with the cathode during electrolysis. cathodic cleaning. Electrolytic cleaning in which the work is the cathode. cathodic corrosion. Corrosion resulting from a cathodic condition of a structure, usually caused by the reaction of an amphoteric metal with the alkaline products of electrolysis. cathodic disbondment. The destruction of adhesion between a coating and its substrate by the products of a cathodic reaction. cathodic inhibitor. A chemical substance or mixture that prevents or reduces the rate of the cathodic or reduction reaction. cathodic pickling. Electrolytic pickling in which the work is the cathode. See also pickling. cathodic polarization. The change of the electrode potential in the active (negative) direction due to current flow. See also polarization. cathodic protection. (1) Reduction of corrosion rate by shifting the potential of the electrode toward a less oxidizing potential with an externally applied voltage. (2) Partial or complete protection of a metal from corrosion by making it a cathode, using either a galvanic cell or an impressed current. Contrast with anodic protection. cathodic reaction. Electrode reaction equivalent to a transfer of negative charge from an electronic to an ionic conductor. A cathodic reaction is a reduction process. A common example in corrosion is: Ox + ne- → Red. catholyte. The electrolyte adjacent to the cathode of an electrolytic cell. cation. A positively charged ion that migrates through the electrolyte toward the cathode under the influence of a potential gradient. See also anion and ion. caustic. (1) Burning or corrosive. (2) A hydroxide of a light metal, such as sodium hydroxide or potassium hydroxide. caustic dip. A strongly alkaline solution into which metal is immersed for etching, for neutralizing acid, or for removing organic materials such as greases or paints. caustic embrittlement or caustic cracking. An obsolete historical term denoting a form of stress-corrosion cracking most frequently encountered in carbon steels or Fe-Cr-Ni alloys that are exposed to concentrated hydroxide solutions at temperatures of 200 to 250 °C (400 to 480 °F). cavitation. The formation and instantaneous collapse of numerous tiny voids or cavities within a liquid caused by rapid and intense pressure changes. Cavitation produced by ultrasonic radiation is sometimes used to

effect violent localized agitation. Cavitation caused by severe turbulent flow often leads to cavitation damage. cavitation corrosion. Material deterioration involving both corrosion and cavitation. cavitation damage. The degradation of a solid body resulting from its exposure to cavitation. This may include loss of the material, surface deformation, or changes in properties or appearance. cavitation-erosion. Progressive loss of the original material from a solid surface due to continuing exposure to cavitation. cell. Electrochemical system consisting of an anode and a cathode immersed in an electrolyte. The anode and cathode may be separate metals or dissimilar areas on the same metal. The cell includes the external circuit, which permits the flow of electrons from the anode to the cathode. See also electrochemical cell. cementite. A hard, brittle compound of iron and carbon, known chemically as iron carbide and having the approximate chemical formula Fe 3C. Characterized by an orthorhombic crystal structure. When it occurs as a phase in steel, the chemical composition will be altered by the presence of manganese and other carbide-forming elements. chalking. The development of loose, removable powder at the surface of an organic coating usually caused by weathering. checking. The development of slight breaks in a coating that do not penetrate to the underlying surface. checks. Numerous, very fine cracks in a coating or at the surface of a metal part. Checks may appear during processing or during service and are most often associated with thermal treatment or thermal cycling. Also called check marks, checking, or heat checks. chelate. (1) A molecular structure in which a heterocyclic ring can be formed by the unshared electrons of neighboring atoms. (2) A coordination compound in which a heterocyclic ring is formed by a metal bound to two atoms of the associated ligand. See also complexation. chelating agent. (1) An organic compound in which atoms form more than one coordinate bond with metals in solution. (2) A substance used in metalfinishing to control or eliminate certain metallic ions present in undesirable quantities. chelation. A chemical process involving formation of a heterocyclic ring compound that contains at least one metal cation or hydrogen ion in the ring. chemical conversion coating. A protective or decorative nonmetallic coating produced in situ by chemical reaction of a metal with a chosen environment. It is often used to prepare the surface prior to the application of an organic coating. chemical potential. In a thermodynamic system of several constituents, the rate of change of the Gibbs function of the system with respect to the change in the number of moles of a particular constituent. chemical vapor deposition (CVD). A coating process, similar to gas carburizing and carbonitriding, whereby a reactant atmosphere gas is fed into a processing chamber where it decomposes at the surface of the workpiece, liberating one material for either absorption by, or accumulation on, the workpiece. A second material is liberated in gas form and is removed from the processing chamber, along with excess atmosphere gas. chemisorption. The binding of an adsorbate to the surface of a solid by forces whose energy levels approximate those of a chemical bond. Contrast with physisorption. chevron pattern.

A fractographic pattern of radial marks (shear ledges) that look like nested letters “V”; sometimes called a herringbone pattern. Typically found on brittle fracture surfaces in structural parts whose widths are considerably greater than their thicknesses. The points of the chevrons can be traced back to the fracture origin. chromadizing. Improving paint adhesion on aluminum or aluminum alloys, mainly aircraft skins, by treatment with a solution of chromic acid. Also called chromodizing or chromatizing. Not to be confused with chromating or chromizing. chromate treatment. A treatment of metal in a solution of a hexavalent chromium compound to produce a conversion coating consisting of trivalent and hexavalent chromium compounds. chromating. Performing a chromate treatment. chrome pickle. (1) Producing a chromate conversion coating on magnesium for temporary protection or for a paint base. (2) The solution that produces the conversion coating. chromizing. A surface treatment at elevated temperature, generally carried out in pack, vapor, or salt bath, in which an alloy is formed by the inward diffusion of chromium into the base metal. cladding. (1) A layer of material, usually metallic, that is mechanically or metallurgically bonded to a substrate. Processes include roll cladding, welding, and explosive forming. (2) A relatively thick (>1 mm, or 0.04 in.) layer of material applied for the purpose of improved corrosion resistance or other properties. clad metal. A composite metal containing two or more layers that have been bonded together. The bonding may have been accomplished by corolling, coextrusion, welding, diffusion bonding, casting, heavy chemical deposition, or heavy electroplating. cleavage. Splitting (fracture) of a crystal on a crystallographic plane of low index. cleavage fracture. A fracture, usually of a polycrystalline metal, in which most of the grains have failed by cleavage, resulting in bright reflecting facets. It is associated with low-energy brittle fracture. cold cracking. A type of weld cracking that usually occurs below 205 °C (400 °F). Cracking may occur during or after cooling to room temperature, sometimes with a considerable time delay. Three factors combine to produce cold cracks: stress (for example, from thermal expansion and contraction), hydrogen (from hydrogen-containing welding consumables), and a susceptible microstructure (plate martensite is most susceptible to cracking, ferritic and bainitic structures least susceptible). See also hot cracking, lamellar tearing, and stress- relief cracking. cold working. Deforming metal plastically under conditions of temperature and strain rate that induce strain hardening. Usually, but not necessarily, conducted at room temperature. Contrast with hot working. combined carbon. The part of the total carbon in steel or cast iron that is present as other than free carbon. complexation. The formation of complex chemical species by the coordination of groups of atoms termed ligands to a central ion, commonly a metal ion. Generally, the ligand coordinates by providing a pair of electrons that forms an ionic or covalent bond to the central ion. See also chelate, coordination compound, and ligand. compressive. Pertaining to forces on a body or part of a body that tend to crush, or compress, the body. compressive strength. The maximum compressive stress a material is capable of developing. With a brittle material that fails in compression by fracturing, the compressive strength has a definite value. In the case of ductile,

malleable, or semiviscous materials (which do not fail in compression by a shattering fracture), the value obtained for compressive strength is an arbitrary value dependent on the degree of distortion that is regarded as effective failure of the material. compressive stress. A stress that causes an elastic body to deform (shorten) in the direction of the applied load. Contrast with tensile stress. concentration cell. An electrochemical cell where the driving force is a difference in concentration of some component in the electrolyte. This difference leads to the formation of discrete cathode and anode sites. concentration polarization. That portion of the polarization of an electrode produced by concentration changes resulting from passage of current through the electrolyte. conductivity. The ratio of the electric current density to the electric field in a material. Also called electrical conductivity or specific conductance. contact corrosion. A term primarily used in Europe to describe galvanic corrosion between dissimilar metals. contact plating. A metal plating process wherein the plating current is provided by galvanic action between the work metal and a second metal, without the use of an external source of current. contact potential. The potential difference at the junction of two dissimilar substances. continuity bond. A metallic connection that provides electrical continuity between metal structures. conversion coating. A coating consisting of a compound formed from the surface metal by chemical or electrochemical treatments. Examples include chromate coatings on zinc, cadmium, magnesium, and aluminum, and oxide and phosphate coatings on steel. See also chromate treatment and phosphating. coordination compound. A compound with a central atom or ion bound to a group of ions or molecules surrounding it. Also called coordination complex. See also chelate, complexation, and ligand. copper-accelerated salt-spray (CASS) test. An accelerated corrosion test for some electrodeposits and for anodic coatings on aluminum. corrodent. A substance that will cause corrosion when brought in contact with a material. A corrosive agent. corrodkote test. An accelerated corrosion test for electrodeposits. corrosion. The chemical or electrochemical reaction between a material and its environment that produces a deterioration of the material and its properties. corrosion effect. A change in any part of the corrosion system caused by corrosion. corrosion embrittlement. The severe loss of ductility of a metal resulting from corrosive attack, usually intergranular and often not visually apparent. corrosion-erosion. See erosion-corrosion. corrosion fatigue. The process in which a metal fractures prematurely under conditions of simultaneous corrosion and repeated cyclic loading at lower stress levels or fewer cycles than would be required in the absence of the corrosive environment. corrosion fatigue strength. The maximum repeated stress that can be endured by metal without failure under definite conditions of corrosion and fatigue and for a specific number of stress cycles and a specified period of time.

corrosion inhibitor. See inhibitor. corrosionist. A person working in the field of corrosion; anyone working to prevent corrosion or protect against corrosion. corrosion potential (Ecorr). The potential of a corroding surface in an electrolyte, relative to a reference electrode. Also called rest potential, open-circuit potential, or freely corroding potential. corrosion product. Substance formed as a result of corrosion. corrosion protection. Modification of a corrosion system so that corrosion damage is mitigated. corrosion rate. Corrosion effect on a metal per unit of time. The type of corrosion rate used depends on the technical system and on the type of corrosion effect. Thus, corrosion rate may be expressed as an increase in the depth of corrosion per unit of time (penetration rate, for example, mils/yr) or the mass of metal turned into corrosion products per unit area of surface per unit of time (weight loss, for example, g/m2/yr). The corrosion effect may vary with time and may not be the same at all points of the corroding surface. Therefore, reports of corrosion rates should be accompanied by information on the type, time dependency, and location of the corrosion effect. corrosion resistance. Ability of a metal to withstand corrosion in a given corrosion system. corrosion system. System consisting of one or more metals and all parts of the environment that influence corrosion. corrosivity. Tendency of an environment to cause corrosion in a given corrosion system. counterelectrode. See auxiliary electrode. couple. See galvanic corrosion. covering power. The ability of a solution to give satisfactory plating at very low current densities, a condition that exists in recesses and pits. This term suggests an ability to cover, but not necessarily to build up, a uniform coating, whereas throwing power suggests the ability to obtain a coating of uniform thickness of an irregularly shaped object. cracking (of coating). Breaks in a coating that extend through to the substrate or underlying surface. crazing. A network of checks or cracks appearing on the surface. creep. Time-dependent strain occurring under stress. The creep strain occurring at a diminishing rate is called primary creep; that occurring at a minimum and almost constant rate, secondary creep; and that occurring at an accelerating rate, tertiary creep. creep-rupture embrittlement. Embrittlement under creep conditions of, for example, aluminum alloys and steels that results in abnormally low rupture ductility. In aluminum alloys, iron in amounts above the solubility limit is known to cause such embrittlement; in steels, the phenomenon is related to the amount of impurities (for example, phosphorus, sulfur, copper, arsenic, antimony, and tin) present. In either case, failure occurs by intergranular cracking of the embrittled material. creep-rupture strength. The stress that will cause fracture in a creep test at a given time in a specified constant environment. Also called stress-rupture strength. crevice corrosion.

Localized corrosion of a metal surface at, or immediately adjacent to, an area that is shielded from full exposure to the environment because of close proximity between the metal and the surface of another material. critical anodic current density. The maximum anodic current density observed during anodic polarization in the active region for a metal or alloy electrode that exhibits active-passive behavior in an environment. critical flaw size. The size of a flaw (defect) in a structure that will cause failure at a particular stress level. critical humidity. The relative humidity above which the atmospheric corrosion rate of some metals increases sharply. critical pitting potential (Ecp, Ep, Epp). The least noble potential at which pits nucleate and grow. It is dependent on the test method used. See also breakdown potential. current (I). The net transfer of electric charge (coulombs) per unit time. The unit of current, the ampere (A), is a base unit of the SI system. See also current density. current density (i). The current flowing to or from a unit area of an electrode surface. Expressed as ampere/meter2 (A/m2) in the SI system. current efficiency. The ratio of the electrochemical equivalent current density for a specific reaction to the total applied current density. D deactivation. The process of prior removal of the active corrosive constituents, usually oxygen, from a corrosive liquid by controlled corrosion of expendable metal or by other chemical means, thereby making the liquid less corrosive. dealloying. The selective corrosion of one or more components of a solid solution alloy. Also called parting or selective leaching. See also dealuminification, decarburization, decobaltification, denickelification, dezincification, and graphitic corrosion. dealuminification. Selective leaching of aluminum, as from aluminum bronze. Also known as dealuminization. decarburization. Loss of carbon from the surface layer of a carbon-containing alloy due to reaction with one or more chemical substances in a medium that contacts the surface. See also dealloying. decobaltification. Corrosion in which cobalt is selectively leached from cobalt-base alloys, such as Stellite, or from cemented carbides. See also dealloying and selective leaching. decomposition potential (or voltage). The potential of a metal surface necessary to decompose the electrolyte in an electrochemical cell or component. deep groundbed. One or more anodes installed vertically at a nominal depth of 15 m (50 ft) or more below the earth's surface in a drilled hole for the purpose of supplying cathodic protection for an underground or submerged metallic structure. See also groundbed. delta ferrite. See ferrite. dendrite. A crystal that has a treelike branching pattern, being most evident in cast metals slowly cooled through the solidification range. denickelification.

Corrosion in which nickel is selectively leached from nickel-containing alloys. Most commonly observed in copper- nickel alloys after extended service in fresh water. See also dealloying and selective leaching. density (of gases). The mass of a unit volume of gas at a stated temperature and pressure. density (of solids and liquids). The mass of unit volume of a material at a specified temperature. deoxidizing. (1) The removal of oxygen from molten metals by use of suitable deoxidizers. (2) Sometimes refers to the removal of undesirable elements other than oxygen by the introduction of elements or compounds that readily react with them. (3) In metalfinishing, the removal of oxide films from metal surfaces by chemical or electrochemical reaction. depolarization. A decrease in the polarization of an electrode depolarizer. A substance that produces depolarization. deposit corrosion. Corrosion occurring under or around a discontinuous deposit on a metallic surface. See also poultice corrosion. descaling. Removing the thick layer of oxides formed on some metals at elevated temperatures. dewetting. The withdrawal of molten solder or zinc from a surface that was previously wetted. If the solderable (or galvanizing) surface is not well protected during soldering (galvanizing), the intermetallic can oxidize and the dewetting phenomenon takes place. dezincification. Corrosion in which zinc is selectively leached from zinc-containing alloys. Most commonly found in copper-zinc alloys containing less than 85% Cu after extended service in water containing dissolved oxygen. See also dealloying and selective leaching. dichromate treatment. A chromate conversion coating produced on magnesium alloys in a boiling solution of sodium dichromate. dielectric shield. In a cathodic protection system, an electrically nonconductive material, such as a coating, plastic sheet, or pipe, that is placed between an anode and an adjacent cathode to avoid current wastage and to improve current distribution, usually on the cathode. differential aeration cell. An electrochemical cell, where the driving force is a difference in the dissolved air (oxygen) concentration in the electrolyte at one electrode compared with that at another electrode of the same material. See also concentration cell. diffusion. (1) Spreading of a constituent in a gas, liquid, or solid, tending to make the composition of all parts uniform. (2) The spontaneous movement of atoms or molecules to new sites within a material. diffusion coating. Any process whereby a base metal or alloy is either (1) coated with another metal or alloy and heated to a sufficient temperature in a suitable environment or (2) exposed to a gaseous or liquid medium containing the other metal or alloy, thus causing diffusion of the coating or of the other metal or alloy into the base metal with resultant changes in the composition and properties of its surface. diffusion coefficient. A factor of proportionality representing the amount of substance diffusing across a unit area through a unit concentration gradient in unit time. diffusion-limited current density. The current density, often referred to as limiting current density, that corresponds to the maximum mass transfer rate that a chemical species can sustain because of diffusion limitations.

dimple rupture. A fractographic term describing ductile fracture that occurs through the formation and coalescence of microvoids along the fracture path. The fracture surface of such a ductile fracture appears dimpled when observed at high magnification and usually is most clearly resolved when viewed in a scanning electron microscope. disbondment. The destruction of adhesion between a coating and the surface coated. discontinuity. Any interruption in the normal physical structure or configuration of a part, such as cracks, laps, seams, inclusions, or porosity. A discontinuity may or may not affect the usefulness of the part. dislocation. A linear imperfection in a crystalline array of atoms. Two basic types are recognized: (1) an edge dislocation corresponds to the row of mismatched atoms along the edge formed by an extra, partial plane of atoms within the body of a crystal; (2) a screw dislocation corresponds to the axis of a spiral structure in a crystal, characterized by a distortion that joins normally parallel planes together to form a continuous helical ramp (with a pitch of one interplanar distance) winding about the dislocation. Most prevalent is the so- called mixed dislocation, which is any combination of an edge dislocation and a screw dislocation. double layer. The interface between an electrode or a suspended particle and an electrolyte created by charge-charge interaction leading to an alignment of oppositely charged ions at the surface of the electrode or particle. The simplest model is represented by a parallel plate condensor. drainage. Conduction of electric current from an underground metallic structure by means of a metallic conductor. Forced drainage is that applied to underground metallic structures by means of an applied electromotive force or sacrificial anode. Natural drainage is that from an underground structure to a more negative (more anodic) structure, such as the negative bus of a trolley substation. dry corrosion. See gaseous corrosion. drying oil. An oil capable of conversion from a liquid to a solid by slow reaction with oxygen in the air. ductile fracture. Fracture characterized by tearing of metal accompanied by appreciable gross plastic deformation and expenditure of considerable energy. Contrast with brittle fracture. ductility. The ability of a material to deform plastically without fracturing, measured by elongation or reduction of area in a tensile test, by height of cupping in an Erichsen test, or by other means. dummy cathode. (1) A cathode, usually corrugated to give variable current densities, that is plated at low current densities to preferentially remove impurities from a plating solution. (2) A substitute cathode that is used during adjustment of operating conditions. dummying. Plating with dummy cathodes. E 885 °F (475 °C) embrittlement. Embrittlement of stainless steels upon extended exposure to temperatures between 400 and 510 °C (750 and 950 °F). Caused by fine, chromium-rich precipitates that segregate at grain boundaries; time at temperature directly influences the amount of segregation. Grain-boundary segregation of the chromium-rich precipitates increases strength and hardness, decreases ductility and toughness, and changes corrosion resistance. Can be reversed by heating above the precipitation range. elastic deformation. A change in dimensions directly proportional to and in phase with an increase or decrease in applied force. elasticity.

The property of a material by virtue of which deformation caused by stress disappears upon removal of the stress. A perfectly elastic body completely recovers its original shape and dimensions after release of stress. elastic limit. The maximum stress that a material is capable of sustaining without any permanent strain (deformation) remaining upon complete release of the stress. elastomer. A natural or synthetic polymer, which at room temperature can be stretched repeatedly to at least twice its original length, and which after removal of the tensile load will immediately and forcibly return to approximately its original length. electrical conductivity. See conductivity. electrical isolation. The condition of being electrically separated from other metallic structures or the environment. electrical resistivity. The electrical resistance offered by a material to the flow of current, times the cross-sectional area of current flow and per unit length of current path; the reciprocal of the conductivity. Also called resistivity or specific resistance. electrochemical admittance. The inverse of electrochemical impedance. electrochemical cell. An electrochemical system consisting of an anode and a cathode in metallic contact and immersed in an electrolyte. The anode and cathode may be different metals or dissimilar areas on the same metal surface. See also cell. electrochemical corrosion. Corrosion that is accompanied by a flow of electrons between cathodic and anodic areas on metallic surfaces. electrochemical equivalent. The weight of an element or group of elements oxidized or reduced at 100% efficiency by the passage of a unit quantity of electricity. Usually expressed as gram-equivalents per coulomb. electrochemical impedance. The frequency-dependent complex-valued proportionality factor (ΔE/ΔI) between the applied potential or current and the response signal. This factor is the total opposition of an electrochemical system to the passage of charge. The value is inversely related to the corrosion rate under certain circumstances. Typical units are Ω or Ω · cm2. electrochemical machining. Controlled metal removal by anodic dissolution. Direct current passes through a flowing conductive solution that separates the workpiece (anode) from the electrode tool (cathode). electrochemical noise. Fluctuations in potential or current, or both, originating from the uncontrolled variations in a corrosion process. electrochemical potential. The partial derivative of the total electrochemical free energy of a constituent with respect to the number of moles of this constituent where all factors are kept constant. It is analogous to the chemical potential of a constituent except that it includes the electric as well as chemical contributions to the free energy. The potential of an electrode in an electrolyte relative to a reference electrode. electrochemical series. A list of elements arranged according to their standard electrode potentials, with “noble” metals such as gold being positive and “active” metals such as zinc being negative. electrode. (1) An electronic conductor used to establish electrical contact with an electrolytic part of a circuit. (2) An electronic conductor in contact with an ionic conductor. electrode polarization.

Change of electrode potential with respect to a reference value. Often the free corrosion potential is used as the reference value. The change may be caused, for example, by the application of an external electrical current or by the addition of an oxidant or reductant. electrodeposition. The deposition of a substance on an electrode by passing electric current through an electrolyte. electrode potential. The potential of an electrode in an electrolyte as measured against a reference electrode. The electrode potential does not include any resistance losses in potential in either the solution or external circuit. It represents the reversible work to move a unit charge from the electrode surface through the solution to the reference electrode. electrode reaction. Interfacial reaction equivalent to a transfer of charge between electronic and ionic conductors. See also anodic reaction and cathodic reaction. electrogalvanizing. The electroplating of zinc upon iron or steel. electrokinetic potential. This potential, sometimes called zeta potential, is a potential difference in the solution caused by residual, unbalanced charge distribution in the adjoining solution, producing a double layer. It differs from the electrode potential in that it occurs exclusively in the solution phase; that is, it represents the reversible work necessary to bring unit charge from infinity in the solution up to the interface in question but not through the interface. electroless plating. A process in which metal ions in a dilute aqueous solution are plated out on a substrate by means of autocatalytic chemical reduction. electrolysis. Production of chemical changes of the electrolyte by the passage of current through an electrochemical cell. electrolyte. (1) A chemical substance or mixture, usually liquid, containing ions that migrate in an electric field. (2) A chemical compound or mixture of compounds that when molten or in solution will conduct an electric current. electrolytic cell. An assembly, consisting of a vessel, electrodes, and an electrolyte, in which electrolysis can be carried out. electrolytic cleaning. A process of removing soil, scale, or corrosion products from a metal surface by subjecting it as an electrode to an electric current in an electrolytic bath. electrolytic protection. See cathodic protection. electromotive force. Electrical potential difference or voltage. This difference could be the result of two dissimilar electrodes in an electrolyte when electrochemical reactions occur. See also thermal electromotive force. electromotive force series (emf series). Same as electrochemical series. electron flow. A movement of electrons in an external circuit between an anode and cathode in a corrosion cell; current flow is in the opposite direction to the electron flow. electroplating. Electrodepositing a metal or alloy in an adherent form on an object serving as a cathode. electropolishing. A technique commonly used to prepare metallographic specimens, in which a high polish is produced by making the specimen the anode in an electrolytic cell, where preferential dissolution at high points smooths the surface. electrotinning.

Electroplating tin on an object. Ellingham diagram. See free-energy diagram. embrittlement. The severe loss of ductility or toughness or both, of a material, usually a metal or alloy. Many forms of embrittlement can lead to brittle fracture. Many forms can occur during thermal treatment or elevatedtemperature service (thermally induced embrittlement). Some of these forms of embrittlement, which affect steels, include blue brittleness, 885 °F (475 °C) embrittlement, quench-age embrittlement, sigmaphase embrittlement, strain-age embrittlement, temper embrittlement, tempered martensite embrittlement, and thermal embrittlement. In addition, steels and other metals and alloys can be embrittled by environmental conditions (environmentally assisted embrittlement). The forms of environmental embrittlement include acid embrittlement, caustic embrittlement, corrosion embrittlement, creep-rupture embrittlement, hydrogen embrittlement, liquid metal induced embrittlement, neutron embrittlement, solder embrittlement, solid metal induced embrittlement, and stress-corrosion cracking. endurance limit. The maximum stress that a material can withstand for an infinitely large number of fatigue cycles. See also fatigue strength. enthalpy (H). The sum of the internal energy of a system plus the product of the system volume multiplied by the pressure exerted on the system by its surroundings. entropy (S). The function of the state of a thermodynamic system whose change in any differential reversible process is equal to the heat absorbed by the system from its surroundings divided by the absolute temperature of the system. environment. The surroundings or conditions (physical, chemical, mechanical) in which a material exists. environmental cracking. Brittle fracture of a normally ductile material in which the corrosive effect of the environment is a causative factor. This general term includes corrosion fatigue, high-temperature hydrogen attack, hydrogen blistering, hydrogen embrittlement, liquid metal induced embrittlement, solid metal induced embrittlement, stress-corrosion cracking, and sulfide stress cracking. The following terms have been used in the past in connection with environment cracking, but are becoming obsolete: caustic embrittlement, delayed fracture, season cracking, static fatigue, stepwise cracking, sulfide corrosion cracking, and sulfide stress-corrosion cracking. See also embrittlement. environmentally assisted embrittlement. See embrittlement. epoxy. Resin formed by the reaction of bisphenol and epichlorohydrin. equilibrium (reversible) potential. The potential of an electrode in an electrolytic solution when the forward rate of a given reaction is exactly equal to the reverse rate. The equilibrium potential can only be defined with respect to a specific electrochemical reaction. erosion. Destruction of metals or other materials by the abrasive action of moving fluids, usually accelerated by the presence of solid particles or matter in suspension. When corrosion occurs simultaneously, the term erosion-corrosion is often used. erosion-corrosion. A material damage involving corrosion and erosion in the presence of a moving corrosive and erosive fluid, leading to the accelerated loss of material. eutectic. (1) An isothermal reversible reaction in which a liquid solution is converted into two or more intimately mixed solids on cooling, the number of solids formed being the same as the number of components in

the system. (2) An alloy having the composition indicated by the eutectic point on an equilibrium diagram. (3) An alloy structure of intermixed solid constituents formed by a eutectic reaction. eutectoid. (1) An isothermal reversible reaction in which a solid solution is converted into two or more intimately mixed solids on cooling, the number of solids formed being the same as the number of components in the system. (2) An alloy having the composition indicated by the eutectoid point on an equilibrium diagram. (3) An alloy structure of intermixed solid constituents formed by a eutectoid reaction. exchange current. When the electrode reactions reach equilibrium in a solution, the rate of anodic dissolution equals the rate of cathodic reduction. This rate is known as the exchange current. exchange current density. The exchange current expressed as a current density. exfoliation. Corrosion that proceeds laterally from the sites of initiation along planes parallel to the surface, generally at grain boundaries, forming corrosion products that force metal away from the body of the material, giving rise to a layered appearance. external circuit. The wires, connectors, measuring devices, current sources, etc., that are used to bring about or measure the desired electrical conditions within the test cell. It is this portion of the cell through which electrons travel. F failure. A general term used to imply that a part in service (1) has become completely inoperable, (2) is still operable but is incapable of satisfactorily performing its intended function, or (3) has deteriorated seriously, to the point that it has become unreliable or unsafe for continued use. false Brinelling. Damage to a solid bearing surface characterized by indentations not caused by plastic deformation due to overload, but thought to be due to other causes such as fretting corrosion. Local spots appear when the protective coating on the metal is broken continually by repeated impacts, usually in the presence of corrosive agents. The term should be avoided when a more precise description is possible. Faraday's constant (F). The product of Avogadro's number times the charge on the electron. F is approximately 96,485 coulombs/ gram-equivalent. Faraday's law. (1) The amount of any substance dissolved or deposited in electrolysis is proportional to the total electric charge passed. (2) The amounts of different substances dissolved or deposited by the passage of the same electric charge are proportional to their equivalent weights. fatigue. The phenomenon leading to fracture under repeated or fluctuating stresses having a maximum value less than the tensile strength of the material. Fatigue fractures are progressive and grow under the action of the fluctuating stress. fatigue crack growth rate (da/dN). The rate of crack extension caused by constant-amplitude fatigue loading, expressed in terms of crack extension per cycle of load application. fatigue life (N). The number of cycles of stress that can be sustained prior to failure under a stated test condition. fatigue limit. The maximum stress that presumably leads to fatigue fracture in a specified number of stress cycles. If the stress is not completely reversed, the value of the mean stress, the minimum stress, or the stress ratio should also be stated. Compare with endurance limit. fatigue strength. The maximum stress that can be sustained for a specified number of cycles without failure, the stress being completely reversed within each cycle unless otherwise stated. ferrite.

(1) A solid solution of one or more elements in body-centered cubic iron. Unless otherwise designated (for instance, as chromium ferrite), the solute is generally assumed to be carbon. On some equilibrium diagrams, there are two ferrite regions separated by an austenite area. The lower area is alpha ferrite; the upper, delta ferrite. If there is no designation, alpha ferrite is assumed. (2) In the field of magnetics, substances having the general formula: M2+O · M3+2O3, the trivalent metal often being iron. filiform corrosion. Corrosion that occurs under some coatings in the form of randomly distributed threadlike filaments. film. A thin, not necessarily visible, layer of material. fish eyes. Areas on a steel fracture surface having a characteristic white, crystalline appearance. flakes. Short, discontinuous internal fissures in wrought metals attributed to stresses produced by localized transformation and decreased solubility of hydrogen during cooling after hot working. In a fracture surface, flakes appear as bright silvery areas; on an etched surface, they appear as short, discontinuous cracks. Also called shatter cracks or snowflakes. flame spraying. Thermal spraying in which a coating material is fed into an oxyfuel gas flame, where it is melted. Compressed gas may or may not be used to atomize the coating material and propel it onto a substrate. foreign structure. Any metallic structure that is not intended as part of a cathodic protection system of interest. fouling. An accumulation of deposits. This term includes accumulation and growth of marine organisms on a submerged metal surface and also includes the accumulation of deposits (usually inorganic) on heat exchanger tubing. fouling organism. Any aquatic organism with a sessile adult stage that attaches to and fouls underwater structures of ships. fractography. Descriptive treatment of fracture, especially in metals, with specific reference to photographs of the fracture surface. Macrofractography involves photographs at low magnification (25×). fracture mechanics. A quantitative analysis for evaluating structural behavior in terms of applied stress, crack length, and specimen or machine component geometry. See also linear elastic fracture mechanics. fracture toughness. A generic term for measures or resistance to extension of a crack. The term is sometimes restricted to results of fracture mechanics tests, which are directly applicable in fracture control. However, the term commonly includes results from simple tests of notched or precracked specimens not based on fracture mechanics analysis. Results from tests of the latter type are often useful for fracture control, based on either service experience or empirical correlations with fracture mechanics tests. See also stressintensity factor. free carbon. The part of the total carbon in steel or cast iron that is present in elemental form as graphite or temper carbon. Contrast with combined carbon. free corrosion potential (Ecorr). Corrosion potential in the absence of net electrical current flowing to or from the metal surface. See also corrosion potential. free energy. See Gibbs free energy. free-energy diagram. A graph of the variation with concentration of the Gibbs free energy at constant pressure and temperature. Called Ellingham diagrams, or Richardson-Jeffes diagrams when nomographs are added. free ferrite.

Ferrite that is formed directly from the decomposition of hypoeutectoid austenite during cooling, without the simultaneous formation of cementite. Also called proeutectoid ferrite. free machining. Pertains to the machining characteristics of an alloy to which one or more ingredients have been introduced to give small broken chips, lower power consumption, better surface finish, and longer tool life; among such additions are sulfur or lead to steel, lead to brass, lead and bismuth to aluminum, and sulfur or selenium to stainless steel. free radical. Any molecule or atom that possesses one unpaired electron. In chemical notation, a free radical is symbolized by a single dot (denoting the odd electron) to the right of the chemical symbol. fretting. A type of wear that occurs between tight-fitting surfaces subjected to cyclic relative motion of extremely small amplitude. Usually, fretting is accompanied by corrosion, especially of the very fine wear debris. fretting corrosion. The accelerated deterioration at the interface between contacting surfaces as the result of corrosion and slight oscillatory movement between the two surfaces. fugacity. A function used as an analog of the partial pressure in applying thermodynamics to real systems; at constant temperature it is proportional to the exponential of the ratio of the chemical potential of a constituent of a system divided by the product of the gas constant and the temperature, and it approaches the partial pressure as the total pressure of the gas approaches zero. furan. Resin formed from reactions involving furfuryl alcohol alone or in combination with other constituents. G galvanic anode. A metal that, because of its relative position in the galvanic or electrochemical series, provides sacrificial protection to metals that are more noble in the series when coupled in an electrolyte. galvanic cell. A cell in which spontaneous chemical change is the source of electrical energy. It usually consists of two dissimilar conductors in contact with each other and with an electrolyte, or of two similar conductors in contact with each other and with dissimilar electrolytes. galvanic corrosion. Accelerated corrosion of a metal because of an electrical contact with a more noble metal or nonmetallic conductor in a corrosive electrolyte. galvanic couple. A pair of dissimilar conductors, commonly metals, in electrical contact. See also galvanic corrosion. galvanic couple potential. See mixed potential. galvanic current. The electric current that flows between metals or conductive nonmetals in a galvanic couple. galvanic series. A list of metals and alloys arranged according to their relative corrosion potentials in a given environment. Compare with electrochemical series. galvanize. To coat a metal surface with zinc using any of various processes. galvanneal. To produce a zinc-iron alloy coating on iron or steel by keeping the coating molten after hot dip galvanizing until the zinc alloys completely with the base metal. galvanodynamic. Referring to a technique where current, continuously varied at a selected rate, is applied to an electrode in an electrolyte. galvanometer.

An instrument for indicating or measuring a small electric current by means of a mechanical motion derived from electromagnetic or electrodynamic forces produced by the current. galvanostaircase. Referring to a galvanostep technique for polarizing an electrode in a series of constant current steps where the time duration and current increments or decrements are equal for each step. galvanostatic. A technique where an electrode is maintained at a constant current in an electrolyte. galvanostep. Refers to a technique in which an electrode is polarized in a series of current increments or decrements. gamma iron. The face-centered cubic form of pure iron, stable from 910 to 1400 °C (1670 to 2550 °F). gaseous corrosion. Corrosion with gases and vapors as the only corrosive agents and without any aqueous phase on the surface of the metal. Also called dry corrosion. gel. (1) A colloidal state comprised of interdispersed solid and liquid, in which the solid particles are themselves interconnected or interlaced in three dimensions. (2) A two-phase colloidal system consisting of a solid and a liquid in more solid form than a sol. general corrosion. See uniform corrosion. Gibbs free energy. The thermodynamic function ΔG = ΔH - TΔS where H is enthalpy, T is absolute temperature, and S is entropy. Also called free energy, free enthalpy, or Gibbs function. glass electrode. A glass membrane electrode used to measure pH or hydrogen-ion activity. grain. An individual crystal in a polycrystalline metal or alloy; it may or may not contain twinned regions and subgrains. grain boundary. A narrow zone in a metal corresponding to the transition from one crystallographic orientation to another, thus separating one grain from another; the atoms in each grain are arranged in an orderly pattern. grain-boundary corrosion. Same as intergranular corrosion. See also interdendritic corrosion. graphitic corrosion. Deterioration of gray cast iron in which the metallic constituents are selectively leached or converted to corrosion products leaving the graphite intact. The term graphitization is commonly used to identify this form of corrosion, but is not recommended because of its use in metallurgy for the decomposition of carbide to graphite. See also dealloying and selective leaching. graphitization. A metallurgical term describing the formation of graphite in iron or steel, usually from decomposition of iron carbide at elevated temperatures. Not recommended as a term to describe graphitic corrosion. green liquor. The liquor resulting from dissolving molten smelt from the kraft recovery furnace in water. See also kraft process and smelt. green rot. A form of high-temperature attack on stainless steels, Ni-Cr alloys, and Ni-Cr-Fe alloys subjected to simultaneous oxidation and carburization. Basically, attack occurs first by precipitation of chromium as chromium carbide, then by oxidation of the carbide particles. groundbed. A buried item, such as junk steel or graphite rods, that serves as the anode for the cathodic protection of pipelines or other buried structures. See also deep groundbed. H half cell.

The electrode and associated electrode reaction comprising half of an electrochemical cell. halogen. Any of the elements of the halogen family, consisting of fluorine, chlorine, bromine, iodine, and astatine. hard chromium. Chromium plated for engineering rather than decorative applications. hardenability. The relative ability of a ferrous alloy to form martensite when quenched from a temperature above the upper critical temperature. Hardenability is commonly measured as the distance below a quenched surface at which the metal exhibits a specific hardness (50 HRC, for example) or a specific percentage of martensite in the microstructure. hardfacing. Depositing filler metal on a surface by welding, spraying, or braze welding to increase resistance to abrasion, erosion, wear, galling, impact, or cavitation damage. hard water. Water that contains certain salts, such as those of calcium or magnesium, which form insoluble deposits in boilers and form precipitates with soap. heat-affected zone (HAZ). That portion of the base metal that was not melted during brazing, cutting, or welding, but whose microstructure and mechanical properties were altered by the heat. heat check. A pattern of parallel surface cracks that are formed by alternate rapid heating and cooling of the extreme surface metal, sometimes found on forging dies and piercing punches. There may be two sets of parallel cracks, one set perpendicular to the other. hematite. (1) An iron mineral crystallizing in the rhombohedral system; the most important ore of iron. (2) An iron oxide, Fe2O3, corresponding to an iron content of approximately 70%. high-temperature hydrogen attack. A loss of strength and ductility of steel by high-temperature reaction of absorbed hydrogen with carbides in the steel resulting in decarburization and internal fissuring. holidays. Discontinuities in a coating (such as porosity, cracks, gaps, and similar flaws) that allow areas of base metal to be exposed to any corrosive environment that contacts the coated surface. hot corrosion. An accelerated corrosion of metal surfaces that results from the combined effect of oxidation and reactions with sulfur compounds and other contaminants, such as chlorides, to form a molten salt on a metal surface that fluxes, destroys, or disrupts the normal protective oxide. See also gaseous corrosion. hot cracking. Caused by the segregation at grain boundaries of low-melting constituents in the weld metal. This can result in grain-boundary tearing under thermal contraction stresses. This can be minimized by the use of low-impurity welding materials and proper joint design. Also called solidification cracking. See also cold cracking, lamellar tearing, and stress-relief cracking. hot dip coating. A metallic coating obtained by dipping the base metal into a molten metal. hot shortness. A tendency for some alloys to separate along grain boundaries when stressed or deformed at temperatures near the melting point. Caused by a low-melting constituent, often present only in minute amounts, that is segregated at grain boundaries. hot working. Deforming metal plastically at such a temperature and strain rate that recrystallization takes place simultaneously with the deformation, thus avoiding any strain hardening. Contrast with cold working. humidity tests. A corrosion test involving exposure of specimens at controlled levels of humidity and temperature. Contrast with salt-fog test.

hydrogen-assisted cracking (HAC). See hydrogen embrittlement. hydrogen-assisted stress-corrosion cracking (HSCC). See hydrogen embrittlement. hydrogen blistering. The formation of blisters on or below a metal surface from excessive internal hydrogen pressure. Hydrogen may be formed during cleaning, plating, corrosion, and so forth. hydrogen damage. A general term for the embrittlement, cracking, blistering, and hydride formation that can occur when hydrogen is present in some metals. hydrogen embrittlement. A process resulting in a decrease in the toughness or ductility of a metal due to the presence of atomic hydrogen. Hydrogen embrittlement has been recognized classically as being of two types. The first, known as internal hydrogen embrittlement, occurs when the hydrogen enters molten metal, which becomes supersaturated with hydrogen immediately after solidification. The second type, environmental hydrogen embrittlement, results from hydrogen being absorbed by solid metals. This can occur during elevated-temperature thermal treatments and in service during electroplating, contact with maintenance chemicals, corrosion reactions, cathodic protection, and operating in high- pressure hydrogen. In the absence of residual stress or external loading, environmental hydrogen embrittlement is manifested in various forms, such as blistering, internal cracking, hydride formation, and reduced ductility. With a tensile stress or stress-intensity factor exceeding a specific threshold, the atomic hydrogen interacts with the metal to induce subcritical crack growth leading to fracture. In the absence of a corrosion reaction (polarized cathodically), the usual term used is hydrogen- assisted cracking (HAC) or hydrogen stress cracking (HSC). In the presence of active corrosion, usually at pits or crevices (polarized anodically), the cracking is generally called stress-corrosion cracking (SCC), but should more properly be called hydrogen-assisted stress-corrosion cracking (HSCC). Thus, HSC and electrochemically anodic SCC can operate separately or in combination (HSCC). In some metals, such as high-strength steels, the mechanism is believed to be all, or nearly all, HSC. The participating mechanism of HSC is not always recognized and may be evaluated under the generic heading of SCC. hydrogen-induced cracking (HIC). Same as hydrogen embrittlement. hydrogen overvoltage. Overvoltage associated with the liberation of hydrogen gas. hydrogen stress cracking (HSC). See hydrogen embrittlement. hydrolysis. (1) Decomposition or alteration of a chemical substance by water. (2) In aqueous solutions of electrolytes, the reactions of cations with water to produce a weak base or of anions to produce a weak acid. hydrophilic. Having an affinity for water. Contrast with hydrophobic. hydrophobic. Lacking an affinity for, repelling, or failing to absorb or adsorb water. Contrast with hydrophilic. hygroscopic. (1) Possessing a marked ability to accelerate the condensation of water vapor; applied to condensation nuclei composed of salts that yield aqueous solutions of a very low equilibrium vapor pressure compared with that of pure water at the same temperature. (2) Pertaining to a substance whose physical characteristics are appreciably altered by effects of water vapor. (3) Pertaining to water absorbed by dry soil minerals from the atmosphere; the amounts depend on the physicochemical character of the surfaces, and increase with rising relative humidity. I immersion plating. Depositing a metallic coating on a metal immersed in a liquid solution, without the aid of an external electric current. Also called dip plating.

immunity. A state of resistance to corrosion or anodic dissolution of a metal caused by thermodynamic stability of the metal. impingement corrosion. A form of erosion-corrosion generally associated with the local impingement of a high-velocity, flowing fluid against a solid surface. impressed current. Direct current supplied by a device employing a power source external to the electrode system of a cathodic protection installation. inclusions. Particles of foreign material in a metallic matrix. The particles are usually compounds (such as oxides, sulfides, or silicates), but may be of any substance that is foreign to (and essentially insoluble in) the matrix. incubation period. A period prior to the detection of corrosion while the metal is in contact with a corrodent. industrial atmosphere. An atmosphere in an area of heavy industry with soot, fly ash, and sulfur compounds as the principal constituents. inert anode. An anode that is insoluble in the electrolyte under the conditions prevailing in the electrochemical cell. inhibitor. A chemical substance or combination of substances that, when present in the environment, prevents or reduces corrosion without significant reaction with the components of the environment. inorganic. Being or composed of matter other than hydrocarbons and their derivatives, or matter that is not of plant or animal origin. Contrast with organic. inorganic zinc-rich paint. Coating containing a zinc powder pigment in an inorganic vehicle. intensiostatic. See galvanostatic. intercrystalline corrosion. See intergranular corrosion. intercrystalline cracking. See intergranular cracking. interdendritic corrosion. Corrosive attack that progresses preferentially along interdendritic paths. This type of attack results from local differences in composition, such as coring commonly encountered in alloy castings. intergranular. Between crystals or grains. Also called intercrystalline. Contrast with transgranular. intergranular corrosion. Corrosion occurring preferentially at grain boundaries, usually with slight or negligible attack on the adjacent grains. Also called intercrystalline corrosion. intergranular cracking. Cracking or fracturing that occurs along the boundaries of grains or crystals in a polycrystalline aggregate. Also called intercrystalline cracking. Contrast with transgranular cracking. intergranular fracture. Brittle fracture of a metal in which the fracture is along the boundaries of grains or crystals that form the metal. Also called intercrystalline fracture. Contrast with transgranular fracture. intergranular stress-corrosion cracking (IGSCC). Stress-corrosion cracking in which the cracking occurs along grain boundaries. intermediate electrode. Same as bipolar electrode. internal oxidation.

The formation of isolated particles of corrosion products beneath the metal surface. This occurs as the result of preferential oxidation of certain alloy constituents by inward diffusion of oxygen, nitrogen, sulfur, and so forth. intumescence. The swelling or bubbling of a coating usually because of heating (term currently used in space and fire protection applications). ion. An atom, or group of atoms, that has gained or lost one or more outer electrons and thus carries an electric charge. Positive ions, or cations, are deficient in outer electrons. Negative ions, or anions, have an excess of outer electrons. ion exchange. The reversible interchange of ions between a liquid and solid, with no substantial structural changes in the solid. iron rot. Deterioration of wood in contact with iron-base alloys. irreversible. Chemical reactions that proceed in a single direction and are not capable of reversal. See also reversible. isocorrosion diagram. A graph or chart that shows constant corrosion behavior with changing solution (environment) composition and temperature. K knife-line attack. Intergranular corrosion of an alloy, usually stabilized stainless steel, along a line adjoining or in contact with a weld after heating into the sensitization temperature range. kraft process. A wood-pulping process in which sodium sulfate is used in the caustic soda pulp-digestion liquor. Also called kraft pulping or sulfate pulping. kurtosis. The extent to which a frequency distribution is concentrated about the mean or peaked. It is sometimes defined as the ratio of the fourth moment of the distribution to the square of the second moment. L lamellar corrosion. See exfoliation. lamellar tearing. Occurs in the base metal adjacent to weldments due to high through- thickness strains introduced by weld metal shrinkage in highly restrained joints. Tearing occurs by decohesion and linking along the working direction of the base metal; cracks usually run roughly parallel to the fusion line and are steplike in appearance. Lamellar tearing can be minimized by designing joints to minimize weld shrinkage stresses and joint restraint. See also cold cracking, hot cracking, and stress-relief cracking. Langelier saturation index. An index calculated from total dissolved solids, calcium concentration, total alkalinity, pH, and solution temperature that shows the tendency of a water solution to precipitate or dissolve calcium carbonate. ledeburite. The eutectic of the iron-carbon system, the constituents of which are austenite and cementite. The austenite decomposes into ferrite and cementite on cooling below Ar1, the temperature at which transformation of austenite to ferrite or ferrite plus cementite is completed during cooling. ligand. The molecule, ion, or group bound to the central atom in a chelate or a coordination compound. limiting current density. The maximum current density that can be used to obtain a desired electrode reaction without undue interference such as from polarization. linear elastic fracture mechanics.

A method of fracture analysis that can determine the stress (or load) required to induce fracture instability in a structure containing a cracklike flaw of known size and shape. See also fracture mechanics and stress-intensity factor. lipophilic. Having an affinity for oil. See also hydrophilic and hydrophobic. liquid metal induced embrittlement (LMIE). Catastrophic brittle failure of a normally ductile metal when in contact with a liquid metal and subsequently stressed in tension. local action. Corrosion due to the action of “local cells,” that is, galvanic cells resulting from inhomogeneities between adjacent areas on a metal surface exposed to an electrolyte. local cell. A galvanic cell resulting from inhomogeneities between areas on a metal surface in an electrolyte. The inhomogeneities may be of physical or chemical nature in either the metal or its environment. localized corrosion. Corrosion at discrete sites, for example, crevice corrosion, pitting corrosion, and stress-corrosion cracking. long-line current. Current that flows through the earth from an anodic to a cathodic area of a continuous metallic structure. Usually used only where the areas are separated by considerable distance and where the current results from concentration-cell action. luggin probe. A small tube or capillary filled with electrolyte, terminating close to the metal surface under study. It is used to provide an ionically conducting path without diffusion between an electrode under study and a reference electrode and to reduce the potential (IR) drop in the potential measurement. Also called a Luggin-Haber capillary. M macroscopic. Visible at magnifications to 25×. macrostructure. The structure of metals as revealed by macroscopic examination of the etched surface of a polished specimen. magnetite. Naturally occurring magnetic oxide of iron (Fe3O4). martensite. Generic term for microstructures formed by diffusionless phase transformation in which the parent and product phases have a specific crystallographic relationship. Characterized by an acicular pattern in the microstructure in both ferrous and nonferrous alloys. In alloys where the solute atoms occupy interstitial positions in the martensitic lattice (such as carbon in iron), the structure is hard and highly strained; but where the solute atoms occupy substitutional positions (such as nickel in iron), the martensite is soft and ductile. The amount of high-temperature phase that transforms to martensite on cooling depends to a large extent on the lowest temperature attained, there being a rather distinct beginning temperature (Ms) and a temperature at which the transformation is essentially complete (Mf). mechanical plating. Plating wherein fine metal powders are peened onto the work by tumbling or other means. metal dusting. Accelerated deterioration of metals in carbonaceous gases at elevated temperatures to form a dustlike corrosion product. metallic glass. An alloy having an amorphous or glassy structure. See also amorphous solid. metallizing. (1) The application of an electrically conductive metallic layer to the surface of nonconductors. (2) The application of metallic coatings by nonelectrolytic procedures such as spraying of molten metal and deposition from the vapor phase.

microbial corrosion. See biological corrosion. microbiologically influenced corrosion (MIC). Corrosion inhibited or accelerated by the presence or activity of microorganisms. Preferred term for the effect that microscopic organisms and their by-products have on electrochemical corrosion of metals and alloys. See also biological corrosion. microscopic. Visible at magnifications above 25×. microstructure. The structure of a prepared surface of a metal as revealed by a microscope at a magnification exceeding 25×. mill scale. The heavy oxide layer formed during hot fabrication or heat treatment of metals. mischmetal. A natural mixture of rare earth elements (atomic numbers 57–71) in metallic form. It contains about 50% Ce, the remainder being principally lanthanum and neodymium. Mischmetal is used as an alloying additive in ferrous alloy to scavenge sulfur, oxygen, and other impurities and in magnesium alloys to improve high-temperature strength. mixed potential. The potential of a material (or materials in a galvanic couple) when two or more electrochemical reactions are occurring. Also called galvanic couple potential. moiety. A portion of a molecule, generally complex, having a characteristic chemical property. molal solution. Concentration of a solution expressed in moles of solute divided by 1000 g of solvent. molar solution. Aqueous solution that contains 1 mole (gram-molecular weight) of solute in 1 L of the solution. mole. One mole is the mass numerically equal (in grams) to the molecular mass (weight) of a substance. It is the amount of substance in a system that contains as many elementary units (Avogadro's number, 6.02 × 1023) as there are atoms of carbon in 0.012 kg of the pure nuclide 12C; the elementary unit must be specified and may be an atom, molecule, ion, electron, photon, or even a specified group of such units. molecular weight. The sum of the atomic weights of the atoms in a molecule. monomer. A molecule, usually an organic compound, having the ability to join with a number of identical molecules to form a polymer. N natural aging. Spontaneous aging of a supersaturated solid solution at room temperature. See also aging. Compare with artificial aging. Nernst equation. An equation that expresses the exact reversible potential of a cell in terms of the activities of products and reactants of the cell reactions. Nernst layer, Nernst thickness. The diffusion layer or the hypothetical thickness of this layer as given by the theory of Nernst. It is defined by: id = nFD[(C0 - C/δ], where id is the diffusion-limited current density, D is the diffusion coefficient, C0 is the concentration at the electrode surface, and δ is the Nernst thickness (0.5 mm in many cases of unstirred aqueous electrolytes). neutron embrittlement. Embrittlement resulting from bombardment with neutrons, usually encountered in metals that have been exposed to a neutron flux in the core of a reactor. In steels, neutron embrittlement is evidenced by a rise in the ductile-to-brittle transition temperature. nitriding.

Introducing nitrogen into the surface layer of a solid ferrous alloy by holding at a suitable temperature (below Ac1 for ferritic steels) in contact with a nitrogenous material, usually ammonia or molten cyanide of appropriate composition. Quenching is not required to produce a hard case. nitrocarburizing. Any of several processes in which both nitrogen and carbon are absorbed into the surface layers of a ferrous material at temperatures below the lower critical temperature and, by diffusion, create a concentration gradient. Performed primarily to provide an antiscuffing surface layer and to improve fatigue resistance. Compare with carbonitriding. noble. The positive direction of electrode potential, thus resembling noble metals such as gold and platinum. noble metal. (1) A metal whose potential is highly positive relative to the hydrogen electrode. (2) A metal with marked resistance to chemical reaction, particularly to oxidation and to solution by inorganic acids. The term as often used is synonymous with precious metal. noble potential. A potential more cathodic (positive) than the standard hydrogen potential. normalizing. Heating a ferrous alloy to a suitable temperature above the transformation range and then cooling in air to a temperature substantially below the transformation range. normal solution. An aqueous solution containing one gram equivalent of the active reagent in 1 L of the solution. normal stress. The stress component perpendicular to a plane on which forces act. Normal stress may be either tensile or compressive. O occluded cell. An electrochemical cell created at a localized site on a metal surface that has been partially obstructed from the bulk environment. open-circuit potential. The potential of an electrode measured with respect to a reference electrode or another electrode when no current flows to or from it through an external circuit. See also corrosion potential. organic. Being or composed of hydrocarbons or their derivatives, or matter of plant or animal origin. Contrast with inorganic. organic acid. A chemical compound with one or more carboxyl radicals (COOH) in its structure; examples are butyric acid, CH3(CH2)2COOH; maleic acid, HOOCCH- CHCOOH; and benzoic acid, C6H5COOH. organic zinc-rich paint. Coating containing zinc powder pigment and an organic resin. overaging. Aging under conditions of time and temperature greater than those required to obtain maximum change in a certain property, so that the property is altered in the direction of the initial value. overheating. Heating a metal or alloy to such a high temperature that its properties are impaired. When the original properties cannot be restored by further heat treating, by mechanical working, or by a combination of working and heat treating, the overheating is known as burning. overvoltage. The difference between the electrode potential when appreciable electrochemical reaction occurs and the reversible electrode potential. oxidation. (1) A reaction in which there is an increase in valence resulting from a loss of electrons. Contrast with reduction. (2) A corrosion reaction in which the corroded metal forms an oxide; usually applied to reaction with a gas containing elemental oxygen, such as air. oxidized surface (on steel).

Surface having a thin, tightly adhering, oxidized skin (from straw to blue in color), extending in from the edge of a coil or sheet. oxidizing agent. A compound that causes oxidation, thereby itself being reduced. oxygen concentration cell. See differential aeration cell. ozone. A powerfully oxidizing allotropic form of the element oxygen. The ozone molecule contains three atoms (O3). Ozone gas is decidedly blue, both liquid and solid ozone are an opaque blue-black color, similar to ink. P partial annealing. An imprecise term used to denote a treatment given cold-worked material to reduce its strength to a controlled level or to effect stress relief. To be meaningful, the type of material, degree of cold work, and the time-temperature schedule must be stated. parting. See dealloying. parts per billion. A measure of proportion by weight, equivalent to one unit weight of a material per billion (109) unit weights of compound. One part per billion is equivalent to 0.001 μg/g or 1 μg/kg. parts per million. A measure of proportion by weight, equivalent to one unit weight of a material per million (106) unit weights of compound. One part per million is equivalent to 1 μg/g or 1 mg/kg. passivation. (1) A reduction of the anodic reaction rate of an electrode involved in corrosion. (2) The process in metal corrosion by which metals become passive. (3) The changing of a chemically active surface of a metal to a much less reactive state. Contrast with activation. passivator. A type of inhibitor that appreciably changes the potential of a metal to a more noble (positive) value. passive. (1) A metal corroding under the control of surface reaction product. (2) The state of the metal surface characterized by low corrosion rates in a potential region that is strongly oxidizing for the metal. passive-active cell. A corrosion cell in which the anode is a metal in the active state, and the cathode is the same metal in the passive state. passivity. A condition in which a piece of metal, because of an impervious covering of oxide or other compound, has a potential much more positive than that of the metal in the active state. patina. The coating, usually green, that forms on the surface of metals such as copper and copper alloys exposed to the atmosphere. Also used to describe the appearance of a weathered surface of any metal. pearlite. A metastable lamellar aggregate of ferrite and cementite resulting from the transformation of austenite at temperatures above the bainite range. pH. The negative logarithm of the hydrogen-ion activity; it denotes the degree of acidity or basicity of a solution. At 25 °C (77 °F), 7.0 is the neutral value. Decreasing values below 7.0 indicate increasing acidity; increasing values above 7.0 indicate increasing basicity. phosphating. Forming an adherent phosphate coating on a metal by immersion in a suitable aqueous phosphate solution. Also called phosphatizing. See also conversion coating. physical vapor deposition (PVD). A coating process whereby deposition species are transferred and deposited in the form of individual atoms or molecules. The most common PVD methods are sputtering and evaporation. Sputtering

involves the transport of a material from a source (target) to a substrate by means of bombardment of the target by gas ions accelerated through a high voltage in a vacuum chamber. In the evaporation process, the transport of a streaming vapor generated by melting and evaporating a coating material source bar by an electron beam in an evacuated chamber, coats the object. physisorption. The binding of an adsorbate to the surface of a solid by forces whose energy levels approximate those of condensation. Contrast with chemisorption. pickle. A solution or process used to loosen or remove corrosion products such as scale or tarnish. pickling. Removing surface oxides from metals by chemical or electrochemical reaction. pitting. Localized corrosion of a metal surface, confined to a point or small area, that takes the form of cavities. pitting factor. Ratio of the depth of the deepest pit resulting from corrosion divided by the average penetration as calculated from weight loss. pitting potential. See critical pitting potential. plane strain. The stress condition in linear elastic fracture mechanics in which there is zero strain in a direction normal to both the axis of applied tensile stress and the direction of crack growth (that is, parallel to the crack front); most nearly achieved in loading thick plates along a direction parallel to the plate surface. In plane strain, the plane of fracture instability is normal to the axis of the principal tensile stress. plane stress. The stress condition in linear elastic fracture mechanics in which the stress in the thickness direction is zero; most nearly achieved in loading very thin sheet along a direction parallel to the surface of the sheet. In plane stress, the plane of fracture instability is inclined 45° to the axis of the principal tensile stress. plasma spraying. A thermal spraying process in which the coating material is melted with heat from a plasma torch that generates a nontransferred arc; molten coating material is propelled against the base metal by the hot, ionized gas issuing from the torch. plastic deformation. The permanent (inelastic) distortion of metals under applied stresses that strain the material beyond its elastic limit. plasticity. The property that enables a material to undergo permanent deformation without rupture. polarization. (1) The change from the open-circuit electrode potential as the result of the passage of current. (2) A change in the potential of the electrodes in an electrolytic cell such that the potential of the anode becomes more noble, and that of the cathode more active, than their respective reversible potentials. Often accompanied by formation of a film on the electrode surface. See also overvoltage. polarization admittance. The reciprocal of polarization resistance (di/dE). polarization curve. A plot of current or current density versus electrode potential for a specific electrode-electrolyte combination. polarization resistance. The slope (dE/di) at the corrosion potential of a potential (E)/current density (i) curve. Also used to describe the method of measuring corrosion rate using this slope. polyester. Resin formed by condensation of polybasic and monobasic acids with polyhydric alcohols. polymer. A chain of organic molecules produced by the joining of primary units called monomers.

potential. Any of various functions from which intensity or velocity at any point in a field may be calculated. The driving influence of an electrochemical reaction. See also active potential, chemical potential, corrosion potential, critical pitting potential, decomposition potential, electrochemical potential, electrode potential, electrokinetic potential, equilibrium (reversible) potential, free corrosion potential, noble potential, open-circuit potential, protective potential, redox potential, and standard electrode potential. potential-pH diagram. See Pourbaix (potential-pH) diagram. potentiodynamic (potentiokinetic). The technique for varying the potential of an electrode in a continuous manner at a preset rate. potentiostat. An instrument for automatically maintaining an electrode in an electrolyte at a constant potential or controlled potentials with respect to a suitable reference electrode. potentiostatic. The technique for maintaining a constant electrode potential. poultice corrosion. A term used in the automotive industry to describe the corrosion of vehicle body parts due to the collection of road salts and debris on ledges and in pockets that are kept moist by weather and washing. Also called deposit corrosion or attack. Pourbaix (potential-pH) diagram. A plot of the redox potential of a corroding system versus the pH of the system, compiled using thermodynamic data and the Nernst equation. The diagram shows regions within which the metal itself or some of its compounds are stable and other regions where the metal corrodes. powder metallurgy. The art of producing metal powders and utilizing metal powders for production of massive materials and shaped objects. precious metal. One of the relatively scarce and valuable metals: gold, silver, and the platinum-group metals. See also noble metal. precipitation hardening. Hardening caused by the precipitation of a constituent from a supersaturated solid solution. See also age hardening and aging. precipitation heat treatment. Artificial aging in which a constituent precipitates from a supersaturated solid solution. precracked specimen. A specimen that is notched and subjected to alternating stresses until a crack has developed at the root of the notch. primary current distribution. The current distribution in an electrolytic cell that is free of polarization. primary passive potential (passivation potential). The potential corresponding to the maximum active current density (critical anodic current density) of an electrode that exhibits active-passive corrosion behavior. primer. The first coat of paint applied to a surface. Formulated to have good bonding and wetting characteristics; may or may not contain inhibiting pigments. principal stress (normal). The maximum or minimum value of the normal stress at a point in a plane considered with respect to all possible orientations of the considered plane. On such principal planes, the shear stress is zero. There are three principal stresses on three mutually perpendicular planes. The state of stress at a point may be (1) uniaxial, a state of stress in which two of the three principal stresses are zero, (2) biaxial, a state of stress in which only one of the three principal stresses is zero, and (3) triaxial, a state of stress in which none of the principal stresses is zero. Multiaxial stress refers to either biaxial or triaxial stress. protection potential (Eprot, Epp). The least noble potential at which existing pits can either passivate or continue growing.

protective potential. The threshold value of the corrosion potential that has to be reached to enter a protective potential range. protective potential range. A range of corrosion potential values in which an acceptable corrosion resistance is achieved for a particular purpose. Q quench-age embrittlement. Embrittlement of low-carbon steels resulting from precipitation of solute carbon at existing dislocations and from precipitation hardening of the steel caused by differences in the solid solubility of carbon in ferrite at different temperatures. Usually caused by rapid cooling of the steel from temperatures slightly below Ac1 (the temperature at which austenite begins to form) and can be minimized by quenching from lower temperatures. quench aging. Aging induced by rapid cooling after solution heat treatment. quench cracking. Fracture of a metal during quenching from elevated temperature. Most frequently observed in hardened carbon steel, alloy steel, or tool steel parts of high hardness and low toughness. Cracks often emanate from fillets, holes, corners, or other stress raisers and result from high stresses due to the volume changes accompanying transformation to martensite. quench hardening. (1) Hardening suitable α-β alloys (most often certain copper or titanium alloys) by solution treating and quenching to develop a martensitelike structure. (2) In ferrous alloys, hardening by austenitizing and then cooling at a rate such that a substantial amount of austenite transforms to martensite. quenching. Rapid cooling of metals (often steels) from a suitable elevated temperature. This generally is accomplished by immersion in water, oil, polymer solution, or salt, although forced air is sometimes used. R radiation damage. A general term for the alteration of properties of a material arising from exposure to ionizing radiation (penetrating radiation), such as x-rays, gamma rays, neutrons, heavy-particle radiation, or fission fragments in nuclear fuel material. rare earth metal. One of the group of 15 chemically similar metals with atomic numbers 57 through 71, commonly referred to as the lanthanides. reactive metal. A metal that readily combines with oxygen at elevated temperatures to form very stable oxides, for example, titanium, zirconium, and beryllium. May also become embrittled by the interstitial absorption of oxygen, hydrogen, and nitrogen. recrystallization. (1) Formation of a new, strain- free grain structure from that existing in cold- worked metal, usually accomplished by heating. (2) The change from one crystal structure to another, as occurs on heating or cooling through a critical temperature. redox potential. The potential of a reversible oxidation-reduction electrode measured with respect to a reference electrode in a given electrolyte. reducing agent. A compound that causes reduction, thereby itself becoming oxidized. reduction. A reaction in which there is a decrease in valence resulting from a gain in electrons. Contrast with oxidation. reference electrode. A nonpolarizable electrode with a known and highly reproducible potential used for potentiometric and voltammetric analyses, for example, the calomel electrode.

refractory metal. A metal having an extremely high melting point, for example, tungsten, molybdenum, tantalum, niobium, chromium, vanadium, and rhenium. In the broad sense, this term refers to metals having melting points above the range for iron, cobalt, and nickel. relative humidity. The ratio, expressed as a percentage, of the amount of water vapor present in a given volume of air at a given temperature to the amount required to saturate the air at that temperature. residual stress. Stresses that remain within a body as a result of plastic deformation. resistance. The opposition that a device or material offers to the flow of direct current, equal to the voltage drop across the element divided by the current through the element. Also called electrical resistance. rest potential. See corrosion potential and open- circuit potential. reversible. A chemical reaction that can proceed in either direction by a change in the system parameters (temperature, pressure, volume, concentration of reactants). Reynold's number. In fluid mechanics, a unitless number, NR that characterizes the flow of liquids. NR = ν · d · ρ/μ where ν is the velocity of the liquid, d is the diameter of the liquid channel, ρ is the density of the liquid, and μ is the viscosity of the liquid. Generally, if NR is less than 2000, the flow is characterized as laminar, with the molecules of the liquid tending to move in straight lines without turbulence. riser. (1) That section of pipeline extending from the ocean floor up the platform. Also, the vertical tube in a steam generator convection bank that circulates water and steam upward. (2) A reservoir of molten metal connected to a casting to provide additional metal to the casting, required as the result of shrinkage before and during solidification. rust. A visible corrosion product consisting of oxides and hydrated oxides of iron. Applied only to ferrous alloys. See also white rust. S sacrificial protection. Reduction of corrosion of a metal in an electrolyte by galvanically coupling it to a more active (or anodic) metal; a form of cathodic protection. salt-fog test. An accelerated corrosion test in which specimens are exposed to a fine mist of a solution usually containing sodium chloride, but sometimes modified with other chemicals. salt-spray test. See salt-fog test. saponification. The alkaline hydrolysis of fats whereby a soap is formed; more generally, hydrolysis of an ester by an alkali with the formation of an alcohol and a salt of the acid portion. saturated calomel electrode. A reference electrode composed of mercury, mercurous chloride (calomel), and a saturated aqueous chloride solution. scale. A solid layer of corrosion products formed on a metal at high temperatures. In some countries the term is also used for deposits from supersaturated water. scaling. (1) The formation at high temperatures of thick corrosion product layers on a metal surface. (2) The deposition of water-insoluble constituents on a metal surface. season cracking. An obsolete historical term usually applied to stress-corrosion cracking of brass. selective leaching.

Corrosion in which one element is preferentially removed from an alloy, leaving a residue (often porous) of the elements that are more resistant to the particular environment. Also called dealloying or parting. See also decarburization, decobaltification, denickelification, dezincification, and graphitic corrosion. sensitization. In austenitic stainless steels, the precipitation of chromium carbides, usually at grain boundaries, on exposure to temperatures of ~550 to 850 °C (~1000 to 1550 °F), leaving the grain boundaries depleted of chromium and therefore susceptible to preferential attack by a corroding (oxidizing) medium. sensitizing heat treatment. A heat treatment, whether accidental, intentional, or incidental (as during welding), that causes precipitation of constituents at grain boundaries, often causing the alloy to become susceptible to intergranular corrosion or intergranular stress-corrosion cracking. See also sensitization. shear. That type of force that causes or tends to cause two contiguous parts of the same body to slide relative to each other in a direction parallel to their plane of contact. shear strength. The stress required to produce fracture in the plane of cross section, the conditions of loading being such that the directions of force and of resistance are parallel and opposite, although their paths are offset a specified minimum amount. The maximum load divided by the original cross-sectional area of a section separated by shear. SI. International system of units, the modern metric system, defined in the document, Le Système International d'Unités (universally abbreviated SI). sigma phase. A hard, brittle, nonmagnetic intermediate phase with a tetragonal crystal structure, containing 30 atoms per unit cell, space group P42/mnm, occurring in many binary and ternary alloys of the transition elements. The composition of this phase in the various systems is not the same, and the phase usually exhibits a wide range in homogeneity. Alloying with a third transition element usually enlarges the field of homogeneity and extends it deep into the ternary section. sigma-phase embrittlement. Embrittlement of iron-chromium alloys (most notably austenitic stainless steels) caused by precipitation at grain boundaries of the hard, brittle intermetallic sigma phase during long periods of exposure to temperatures between ~560 and 980 °C (~1050 and 1800 °F). Sigma-phase embrittlement results in severe loss in toughness and ductility and can make the embrittled material susceptible to intergranular corrosion. See also sensitization. slip. Plastic deformation by the irreversible shear displacement (translation) of one part of a crystal relative to another in a definite crystallographic direction and usually on a specific crystallographic plane. Sometimes called glide. slow-strain-rate technique. An experimental technique for evaluating susceptibility to stress-corrosion cracking. It involves pulling the specimen to failure in uniaxial tension at a controlled slow strain rate while the specimen is in the test environment and examining the specimen for evidence of stress-corrosion cracking. smelt. Molten slag; in the pulp and paper industry, the cooking chemicals tapped from the recovery boiler as molten material and dissolved in the smelt tank as green liquor. S-N diagram. A plot showing the relationship of stress, S, and the number of cycles, N, before fracture in fatigue testing. soft water. Water that is free of magnesium or calcium salts. sol. A colloidal suspension comprising discrete or separate solid particles suspended in a liquid. Differs from a solution, though one merges into the other. Compare with gel.

solder embrittlement. Reduction in mechanical properties of a metal as a result of local penetration of solder along grain boundaries. sol-gel process. An important ceramic and glass forming process in which a sol is converted to a gel by particle evaporation of the liquid phase and/or by neutralizing the electric charges on particles that cause them to repel each other. The gel is usually further processed (formed, dried, and fired). solid metal induced embrittlement (SMIE). The occurrence of embrittlement in a material below the melting point of the embrittling species. See also liquid metal induced embrittlement (LMIE). solid solution. A single, solid, homogeneous crystalline phase containing two or more chemical species. solute. The component of either a liquid or solid solution that is present to a lesser or minor extent; the component that is dissolved in the solvent. solution. In chemistry, a homogeneous dispersion of two or more kinds of molecular or ionic species. Solution may be composed of any combination of liquids, solids, or gases, but they are always a single phase. solution heat treatment. Heating an alloy to a suitable temperature, holding at that temperature long enough to cause one or more constituents to enter into solid solution, and then cooling rapidly enough to hold these constituents in solution. solution potential. Electrode potential where half-cell reaction involves only the metal electrode and its ion. solvent. The component of either a liquid or solid solution that is present to a greater or major extent; the component that dissolves the solute. sour gas. A gaseous environment containing hydrogen sulfide and carbon dioxide in hydrocarbon reservoirs. Prolonged exposure to sour gas can lead to hydrogen damage, sulfide stress cracking, and/or stresscorrosion cracking in ferrous alloys. sour water. Waste waters containing fetid materials, usually sulfur compounds. spalling. The spontaneous chipping, fragmentation, or separation of a surface or surface coating. spheroidite. An aggregate of iron or alloy carbides of essentially spherical shape dispersed throughout a matrix of ferrite. sputtering. A coating process whereby thermally emitted electrons collide with inert gas atoms, which accelerate toward and impact a negatively charged electrode that is a target of the coating material. The impacting ions dislodge atoms of the target material, which are in turn projected to and deposited on the substrate to form the coating. stabilizing treatment. (1) Before finishing to final dimensions, repeatedly heating a ferrous or nonferrous part to or slightly above its normal operating temperature and then cooling to room temperature to ensure dimensional stability in service. (2) Transforming retained austenite in quenched hardenable steels, usually by cold treatment. (3) Heating a solution- treated stabilized grade of an austenitic stainless steel to 870 to 900 °C (1600 to 1650 °F) to precipitate all carbon as TiC, NbC, or TaC so that sensitization is avoided on subsequent exposure to elevated temperature. standard electrode potential. The reversible potential for an electrode process when all products and reactions are at unit activity on a scale in which the potential for the standard hydrogen half-cell is zero. strain.

The unit of change in the size or shape of a body due to force. Also known as nominal strain. strain-age embrittlement. A loss in ductility accompanied by an increase in hardness and strength that occurs when low-carbon steel (especially rimmed or capped steel) is aged following plastic deformation. The degree of embrittlement is a function of aging time and temperature, occurring in a matter of minutes at about 200 °C (400 °F), but requiring a few hours to a year at room temperature. strain aging. Aging induced by cold working. strain hardening. An increase in hardness and strength caused by plastic deformation at temperatures below the recrystallization range. strain rate. The time rate of straining for the usual tensile test. Strain as measured directly on the specimen gage length is used for determining strain rate. Because strain is dimensionless, the units of strain rate are reciprocal time. stray current. Current flowing underground or in soils along paths other than the intended circuit. stray-current corrosion. Corrosion resulting from direct current flow along paths other than the intended circuit. stress. The intensity of the internally distributed forces or components of forces that resist a change in the volume or shape of a material that is or has been subjected to external forces. Stress is expressed in force per unit area and is calculated on the basis of the original dimensions of the cross section of the specimen. Stress can be either direct (tension or compression) or shear. See also residual stress. stress concentration factor (Kt). A multiplying factor for applied stress that allows for the presence of a structural discontinuity such as a notch or hole; Kt equals the ratio of the greatest stress in the region of the discontinuity to the nominal stress for the entire section. Also called theoretical stress concentration factor. stress-corrosion cracking (SCC). A cracking process that requires the simultaneous action of a corrodent and sustained tensile stress. This excludes corrosion-reduced sections that fail by fast fracture. It also excludes intercrystalline or transcrystalline corrosion, which can disintegrate an alloy without applied or residual stress. Stresscorrosion cracking may occur in combination with hydrogen embrittlement. stress-intensity factor. A scaling factor, usually denoted by the K, used in linear elastic fracture mechanics to describe the intensification of applied stress at the tip of a crack of known size and shape. At the onset of rapid crack propagation in any structure containing a crack, the factor is called the critical stress- intensity factor, or the fracture toughness. Various subscripts are used to denote different loading conditions or fracture toughnesses. Kc. Plane-stress fracture toughness. The value of stress intensity at which crack propagation becomes rapid in sections thinner than those in which plane-strain conditions prevail. KI. Stress-intensity factor for a loading condition that displaces the crack faces in a direction normal to the crack plane (also known as the opening mode of deformation). KIc. Plane-strain fracture toughness. The minimum value of Kc for any given material and condition, which is attained when rapid crack propagation in the opening mode is governed by plane-strain conditions. KId. Dynamic fracture toughness. The fracture toughness determined under dynamic loading conditions; it is used as an approximation of KIc for very tough materials. KISCC. Threshold stress-intensity factor for stress-corrosion cracking. The critical plane-strain stress intensity at the onset of stress-corrosion cracking under specified conditions.

KQ. Provisional value for plane-strain fracture toughness. Kth. Threshold stress intensity for stress-corrosion cracking. The critical stress intensity at the onset of stresscorrosion cracking under specified conditions. ΔK. The range of the stress-intensity factor during a fatigue cycle, that is, Kmax - Kmin. stress-raisers. Changes in contour or discontinuities in structure that cause local increases in stress. stress ratio (A or R). The algebraic ratio of two specified stress values in a stress cycle. Two commonly used stress ratios are: (1) the ratio of the alternating stress amplitude to the mean stress, A = Sa/Sm; (2) the ratio of the minimum stress to the maximum stress, R = Smin/ Smax. stress-relief cracking. Occurs when susceptible alloys are subjected to thermal stress relief after welding to reduce residual stresses and improve toughness. Occurs only in metals that can precipitation harden during such elevated- temperature exposure; it usually occurs at stress raisers, is intergranular in nature, and is generally observed in the coarse-grained region of the weld heat-affected zone. Also called postweld heat treatment cracking. See also cold cracking, hot cracking, and lamellar tearing. striation. A fatigue fracture feature, often observed in electron micrographs, that indicates the position of the crack front after each succeeding cycle of stress. The distance between striations indicates the advance of the crack front across that crystal during one stress cycle, and a line normal to the striations indicates the direction of local crack propagation. See also beach marks. subsurface corrosion. See internal oxidation. sulfidation. The reaction of a metal or alloy with a sulfur-containing species to produce a sulfur compound that forms on or beneath the surface on the metal or alloy. sulfide stress cracking. Brittle failure by cracking under the combined action of tensile stress and corrosion in the presence of water and hydrogen sulfide. See also environmental cracking. surface profile. Anchor pattern on a surface produced by abrasive blasting or acid treatment. surfactant. A surface-active agent; usually an organic compound whose molecules contain a hydrophilic group at one end and a lipophilic group at the other. T Tafel line, Tafel slope, Tafel diagram. When an electrode is polarized, it frequently will yield a current/potential relationship over a region that can be approximated by: η = ±B log (i/ io), where η is the change in open-circuit potential, i is the current density, and B and io are constants. The constant B is also known as the Tafel slope. If this behavior is observed, a plot on semilogarithmic coordinates yields a straight line known as the Tafel line, and the overall diagram is termed a Tafel diagram. tarnish. Surface discoloration of a metal caused by formation of a thin film of corrosion product. temper. (1) In heat treatment, to reheat hardened steel or hardened cast iron to some temperature below the eutectoid temperature for the purpose of decreasing hardness and increasing toughness. The process is also sometimes applied to normalized steel. (2) In tool steels, temper is sometimes inadvisably used to denote carbon content. (3) In nonferrous alloys and in some ferrous alloys (steels that cannot be hardened by heat treatment), the hardness and strength produced by mechanical or thermal treatment, or both, and characterized by a certain structure, mechanical properties, or reduction of area during cold working.

temper color. A thin, tightly adhering oxide skin (only a few molecules thick) that forms when steel is tempered at a low temperature, or for a short time, in air or a mildly oxidizing atmosphere. The color, which ranges from straw to blue depending on the thickness of the oxide skin, varies with both tempering time and temperature. tempered martensite embrittlement. Embrittlement of ultrahigh-strength steels caused by tempering in the temperature range of 205 to 400 °C (400 to 750 °F); also called 350 °C or 500 °F embrittlement. Tempered martensite embrittlement is thought to result from the combined effects of cementite precipitation on prior-austenite grain boundaries or interlath boundaries and the segregation of impurities at prior-austenite grain boundaries. temper embrittlement. Embrittlement of alloy steels caused by holding within or cooling slowly through a temperature range just below the transformation range. Embrittlement is the result of the segregation at grain boundaries of impurities such as arsenic, antimony, phosphorus, and tin; it is usually manifested as an upward shift in ductile-to-brittle transition temperature. Temper embrittlement can be reversed by retempering above the critical temperature range, then cooling rapidly. tensile strength. In tensile testing, the ratio of maximum load to original cross-sectional area. Also called ultimate tensile strength. tensile stress. A stress that causes two parts of an elastic body, on either side of a typical stress plane, to pull apart. Contrast with compressive stress. tension. The force or load that produces elongation. terne. An alloy lead containing 3 to 15% Sn, used as a hot dip coating for steel sheet or plate. Terne coatings, which are smooth and dull in appearance, give the steel better corrosion resistance and enhance its ability to be formed, soldered, or painted. thermal electromotive force. The electromotive force generated in a circuit containing two dissimilar metals when one junction is at a temperature different from that of the other. See also electromotive force and thermocouple. thermal embrittlement. Intergranular fracture of maraging steels with decreased toughness resulting from improper processing after hot working. Occurs upon heating above 1095 °C (2000 °F) and then slow cooling through the temperature range of 815 to 980 °C (1500 to 1800 °F). Has been attributed to precipitation of titanium carbides and titanium carbonitrides at austenite grain boundaries during cooling through the critical temperature range. thermally induced embrittlement. See embrittlement. thermal spraying. A group of coating or welding processes in which finely divided metallic or nonmetallic materials are deposited in a molten or semimolten condition to form a coating. The coating material may be in the form of powder, ceramic rod, wire, or molten materials. See also flame spraying and plasma spraying. thermocouple. A device for measuring temperatures, consisting of lengths of two dissimilar metals or alloys that are electrically joined at one end and connected to a voltage-measuring instrument at the other end. When one junction is hotter than the other, a thermal electromotive force is produced that is roughly proportional to the difference in temperature between the hot and cold junctions. thermogalvanic corrosion. Corrosion resulting from an electrochemical cell caused by a thermal gradient. threshold stress. Threshold stress for stress-corrosion cracking. The critical gross section stress at the onset of stresscorrosion cracking under specified conditions. throwing power.

(1) The relationship between the current density at a point on a surface and its distance from the counterelectrode. The greater the ratio of the surface resistivity shown by the electrode reaction to the volume resistivity of the electrolyte, the better is the throwing power of the process. (2) The ability of a plating solution to produce a uniform metal distribution on an irregularly shaped cathode. Compare with covering power. tinning. Coating metal with a very thin layer of molten solder or brazing filler metal. torsion. A twisting deformation of a solid body about an axis in which lines that were initially parallel to the axis become helices. torsional stress. The shear stress on a transverse cross section resulting from a twisting action. total carbon. The sum of the free carbon and combined carbon (including carbon in solution) in a ferrous alloy. toughness. The ability of a metal to absorb energy and deform plastically before fracturing. transcrystalline. See transgranular. transcrystalline cracking. See transgranular cracking. transference. The movement of ions through the electrolyte associated with the passage of an electric current. Also called transport or migration. transgranular. Through or across crystals or grains. Also called intracrystalline or transcrystalline. transgranular cracking. Cracking or fracturing that occurs through or across a crystal or grain. Also called transcrystalline cracking. Contrast with intergranular cracking. transgranular fracture. Fracture through or across the crystals or grains of a metal. Also called transcrystalline fracture or intracrystalline fracture. Contrast with intergranular fracture. transition metal. A metal in which the available electron energy levels are occupied in such a way that the d-band contains less than its maximum number of ten electrons per atom, for example, iron, cobalt, nickel, and tungsten. The distinctive properties of the transition metals result from the incompletely filled d- levels. transition temperature. (1) An arbitrarily defined temperature that lies within the temperature range in which metal fracture characteristics (as usually determined by tests of notched specimens) change rapidly, such as from primarily fibrous (shear) to primarily crystalline (cleavage) fracture. (2) Sometimes used to denote an arbitrarily defined temperature within a range in which the ductility changes rapidly with temperature. transpassive region. The region of an anodic polarization curve, noble to and above the passive potential range, in which there is a significant increase in current density (increased metal dissolution) as the potential becomes more positive (noble). transpassive state. (1) State of anodically passivated metal characterized by a considerable increase of the corrosion current when the potential is increased. (2) The noble region of potential where an electrode exhibits a current density higher than the passive current density. triaxial stress. See principal stress (normal). tuberculation. The formation of localized corrosion products scattered over the surface in the form of knoblike mounds called tubercles. U

ultimate strength. The maximum stress (tensile, compressive, or shear) a material can sustain without fracture, determined by dividing maximum load by the original cross-sectional area of the specimen. Also called nominal strength or maximum strength. ultramicrotome. An instrument for cutting very thin specimens for study with a microscope. underfilm corrosion. Corrosion that occurs under organic films in the form of randomly distributed threadlike filaments or spots. In many cases this is identical to filiform corrosion. uniaxial stress. See principal stress (normal). uniform corrosion. (1) A type of corrosion attack (deterioration) uniformly distributed over a metal surface. (2) Corrosion that proceeds at approximately the same rate over a metal surface. Also called general corrosion. V vacuum deposition. Condensation of thin metal coatings on the cool surface of work in a vacuum. valence. A positive number that characterizes the combining power of an element for other elements, as measured by the number of bonds to other atoms that one atom of the given element forms upon chemical combination; hydrogen is assigned valence 1, and the valence is the number of hydrogen atoms, or their equivalent, with which an atom of the given element combines. vapor deposition. See chemical vapor deposition, physical vapor deposition, and sputtering. vapor plating. Deposition of a metal or compound on a heated surface by reduction or decomposition of a volatile compound at a temperature below the melting points of the deposit and the base material. The reduction is usually accomplished by a gaseous reducing agent such as hydrogen. The decomposition process may involve thermal dissociation or reaction with the base material. Occasionally used to designate deposition on cold surfaces by vacuum evaporation. See also vacuum deposition. voids. A term generally applied to paints to describe holidays, holes, and skips in a film. Also used to describe shrinkage in castings and welds. W wash primer. A thin, inhibiting paint, usually chromate pigmented with a polyvinyl butyrate binder. weld cracking. Cracking that occurs in the weld metal. See also cold cracking, hot cracking, lamellar tearing, and stress-relief cracking. weld decay. A nonpreferred term for intergranular corrosion, usually of stainless steels or certain nickel-base alloys, that occurs as the result of sensitization in the heat-affected zone during the welding operation. wetting. A condition in which the interfacial tension between a liquid and a solid is such that the contact angle is 0 to 90°. wetting agent. A substance that reduces the surface tension of a liquid, thereby causing it to spread more readily on a solid surface. white liquor. Cooking liquor from the kraft pulping process produced by recausticizing green liquor with lime. white rust. Zinc oxide; the powdery corrosion product on zinc or zinc-coated surfaces. work hardening. Same as strain hardening.

working electrode. The test or specimen electrode in an electrochemical cell. Y yield. Evidence of plastic deformation in structural materials. Also called plastic flow or creep. yield point. The first stress in a material, usually less than the maximum attainable stress, at which an increase in strain occurs without an increase in stress. Only certain metals—those that exhibit a localized, heterogeneous type of transition from elastic deformation to plastic deformation—produce a yield point. If there is a decrease in stress after yielding, a distinction may be made between upper and lower yield points. The load at which a sudden drop in the flow curve occurs is called the upper yield point. The constant load shown on the flow curve is the lower yield point. yield strength. The stress at which a material exhibits a specified deviation from proportionality of stress and strain. An offset of 0.2% is used for many metals. yield stress. The stress level in a material at or above the yield strength, but below the ultimate strength, that is, a stress in the plastic range. Z zeta potential. See electrokinetic potential. Glossary of Terms

Selected References • • • • • • • • • • • • • • • •

ASM Materials Engineering Dictionary, ASM International, 1992 Compilation of ASTM Standard Definitions, 5th ed., American Society for Testing and Materials, 1982 Concise Encyclopedia of Science and Technology, McGraw-Hill, 1984 “Corrosion of Metals and Alloys—Basic Terms and Definitions,” ISO 8044, International Organization for Standardization, 1999 (available from the American National Standards Institute) Dictionary of Scientific and Technical Terms, 5th ed., McGraw-Hill, 1994 Electroplating Engineering Handbook, 3rd ed., A.K. Grahm, Ed., Van Nostrand Reinhold, 1971, p ix– xviii A.D. Merriman, A Dictionary of Metallurgy, Pitman Publishing, 1958 Metals Handbook Desk Edition, American Society for Metals, 1998, p 1–63 Military Standardization Handbook, Corrosion and Corrosion Prevention of Metals, MIL-HDBK-729, Section 10, Department of Defense “NACE Glossary of Corrosion Related Terms,” National Association of Corrosion Engineers, 1985 Science and Technology of Surface Coating, B.N. Chapman and J.C. Anderson, Ed., Academic Press, 1974, p 435–445 “Standard Terminology Relating to Corrosion and Corrosion Testing,” G 15, Annual Book of ASTM Standards, Vol 03.02, ASTM International “Standard Terminology Relating to Electroplating,” B 374, Annual Book of ASTM Standards, Vol 02.05, ASTM International “Terminology Relating to Erosion and Wear,” G 40, Annual Book of ASTM Standards, Vol 03.02, ASTM International “Terminology Relating to Fatigue and Fracture Testing,” E 1823, Annual Book of ASTM Standards, Vol 03.01, ASTM International “Terminology Relating to Methods of Mechanical Testing,” E 6, Annual Book of ASTM Standards, Vol 03.01, ASTM International

Abbreviations and Symbols a crack length; chemical activity A ampere A area Aa anodic area Ac cathodic area Å angstrom AA Aluminum Association ac alternating current ACI Alloy Casting Institute Acorr corroding area A/D-D/A analog-to-digital and digital-to-analog AE auxiliary electrode AES Auger electron spectroscopy AFM atomic force microscopy AISI American Iron and Steel Institute AMS Aerospace Material Specification (of SAE International) amu atomic mass units ANSI American National Standards Institute API American Petroleum Institute AS artificial seawater ASME American Society of Mechanical Engineers ASTM American Society for Testing and Materials (Now ASTM International) at.% atomic percent atm atmospheres (pressure) AW atomic weight AWS American Welding Society b Tafel coefficient bal balance or remainder bcc body-centered cubic bct body-centered tetrragonal BFPD barrels of fluid produced per day (oil and gas production) c capacitance; concentration C coulomb Co initial coating thickness CAD/CAE computer-aided design/computer- aided engineering CASS copper accelerated acetic acid salt spray (test) CCT critical crevice temperature CE counter electrode CCT critical crevice temperature CFC corrosion-fatigue cracking cm centimeter CP commercially pure cpm cycles per minute CPT critical pitting temperature CSE saturated copper-sulfate reference electrode CTE coefficient of thermal expansion CVD chemical vapor deposition CW cold work

d density; used in mathematical expressions involving a derivative (denotes rate of change) D diffusion coefficient da/dN crack growth rate per cycle da/dt crack growth rate per unit time dc direct current e natural log base, 2.71828… e- electron E electrical potential Eo standard potential value Eappl applied potential Ecell reversible electrode cell potential Ecorr corrosion potential Ee equilibrium potential Eg galvanic potential Ep pitting potential; passivation potential Epass passivation potential Epit critical potential for pitting ER repassivation potential EAC environmentally assisted cracking ECM electrochemically machining ECN electrochemical noise (method) EDM electrical discharge machining EDS energy dispersive spectroscopy; energydispersive x-ray spectroscopy EDTA ethylenediamine tetraacetic acid EIS electrochemical impedance spectroscopy emf electromotive force EN electrochemical noise ENA electrochemical noise analysis EPA Environmental Protection Agency (U.S.) EPDM ethylene-propylene-diene monomer (rubber) Eq equation (also used to label inequalities and reactions) ER electrical resistance EVD electrochemical vapor deposition f flow rate F farad F Faraday constant FAA Federal Aviation Administration (U.S.) FACT Ford anodized aluminum corrosion test fcc face-centered cubic FEA finite-element analysis FEM field-emission microscopy Fig. figure FFT fast Fourier transform FGD flue gas desulfurization FRA frequency response analyzer FRP fiber-reinforced polyester ft foot FTIR Fourier transform infrared (spectroscopy) g gas; gram G Gibbs energy gal gallon GNP gross national product GPS global positioning system GTA gas tungsten arc

GTAW gas tungsten arc welded h hour h· electron hole H enthalpy H0 null hypothesis HAZ heat-affected zone HB Brinell hardness hcp hexagonal close-packed HIC hydrogen-induced cracking HR Rockwell hardness; requires scale designation, such as HRC for Rockwell C hardness HV Vickers (diamond pyramid) hardness Hz hertz i current density; current icorr corrosion current (density) icrit critical current for passivation ip passive current density I current; current density Iappl applied current Icorr corrosion current IC integrated circuit ICCP impressed current cathodic protection ID inside diameter IGA intergranular attack IGC intergranular corrosion in. inch ipy inches per year ISO International Organization for Standardization IR voltage drop, current multiplied by resistance IUPAC International Union of Pure and Applied Chemistry j , current density J flux or mass; stress-intensity factor in elastic- plastic fracture mechanics; current density Jo exchange current density kB Boltzmann constant, 1.38066 × 10-23 J/K kL linear oxidation rate constant kp parabolic oxidation rate constant K Kelvin K stress-intensity factor in linear elastic fracture mechanics Kcrit critical value of stress concentration KIc fracture toughness; plane-strain fracture toughness KISCC threshold stress intensity for stress-corrosion cracking Kt stress-concentration factor Kth threshold stress-intensity factor kg kilogram kPa kilopascal l liquid L liter lb pound LME liquid metal embrittlement LMIE liquid metal induced embrittlement ln natural logarithm (base e) log common logarithm (base 10) LPR linear polarization resistance LSI Langelier saturation index m mass; molar (solution)

mc mass transport coefficient M metal M molecular weight; molar solution mA milliampere max maximum MCA multiple-crevice assembly mdd milligrams per square decimeter per day MEM maximum entropy method mg milligram MIC microbiologically influenced corrosion MICI microbiologically influenced corrosion inhibition mil 0.001 in. MIL-STD military standard (U.S.) min minimum; minute mm millimeter mol mole MPa megapascal mpy mils per year mV millivolt n sample size N Newton N population size, number of trials; mole fraction; normal (solution); number of cycles in corrosion-fatigue testing NACE National Association of Corrosion Engineers (now NACE-International) NIST National Institute of Standards and Technology nm nanometer No. number NSF National Sanitation Foundation OD outside diameter OEM original equipment manufacturer OSHA Occupational Safety and Health Administration (U.S.) p page p probability of success; partial pressure; pressure pO2 partial pressure of oxygen P probability Pa pascal Pa sensitization number PASS paint adhesion on a scribed surface pH negative logarithm of hydrogen ion activity ppb parts per billion ppm parts per million PREN pitting-resistance equivalent number PSD power spectral density psi pounds per square inch psia pounds per square inch (absolute) psig pounds per square inch (gage) PTFE polytetrafluoroethylene PVC polyvinyl chloride PVD plasma vapor deposition Q activation energy for diffusion, heat r number of successes; corrosion rate R Rankine; reduction; reduced species R radius; ration of minimum stress to maximum stress; gas constant; resistance Ri internal resistance

Rn noise resistance Rp polarization resistance RE reference electrode Re Reynold's number Ref reference RF radiofrequency RH relative humidity rms root mean square s estimate of the standard deviation S entropy Sa stress amplitude Sc Schmidt number Sm mean stress Sr fatigue (endurance) limit; stress range SAE Society of Automotive Engineers SCC stress-corrosion cracking SCE saturated calomel electrode SCR silicon-controlled rectifier SEM scanning electron microscopy SERS surface enhanced Raman spectroscopy Sh Sherwood number SHE standard hydrogen electrode SLPR self-linear polarization resistance SMIE solid metal induced embrittlement SRB sulfate-reducing bacteria; solid rocket booster (space shuttle) SSPC The Society for Protective Coatings; Steel Structures Painting Council t thickness; time T temperature Tg glass transition temperature TDS total dissolved solids TFE tetrafluoroethylene TGA thermogravimetric analysis TOW time of wetness UNS Unified Numbering System UTS ultimate tensile strength UV ultraviolet V volt V volume v viscosity VNSS Väätänen nine salts solution VOC volatile organic compound vol% volume percent w weight or mass W watt WDS wavelength dispersive spectroscopy WE working electrode wt% weight percent yr year Z impedance ZRA zero-resistance ammetry; zero-resistance ammeter ° degree (angular measure) °C temperature, degrees Celsius (centigrade) °F temperature, degrees Fahrenheit → chemical reaction

↔ reversible chemical reaction, does not imply equal reaction rates in both directions reversible chemical reaction = equals ≈ approximately equals ≠ not equal to > greater than » much greater than ≥ greater than or equal to ∫ integral < less than ≤ less than or equal to ± maximum deviation. tolerance - minus; negative ion charge × multiplied by; diameters (magnification) · multiplied by / per % percent + plus; in addition to; positive ion charge √ square root of ~ similar to; approximately α varies as; is proportional to α chemical activity; crack length; crystal lattice length along the α axis β Tafel coefficient Δ change in quantity, an increment, a range ΔEtherm thermodynamic driving force ΔG Gibbs free energy ΔG0 standard Gibbs free energy ΔH change in enthalpy ΔK stress-intensity factor range ΔS entropy change εp plastic strain range εt total strain range η overpotential μm micron (micrometer) ν kinematic viscosity Φ phase angle Π symbol for multiplying a series of terms π pi, 3.14159… ρ density σ, σ′ standard deviation; oxide conductivity; stress τ0 time constant Σ summation Ω ohm ω angular velocity Greek Alphabet Α, α alpha Β, β beta Γ, γ gamma Δ, δ delta Ε, ε epsilon Ζ, ζ zeta Η, η eta Θ, θ theta Ι, ι iota

Κ, κ Λ, λ Μ, μ Ν, ν Ξ, ξ Ο, ο Π, π Ρ, ρ Σ, σ Τ, τ Υ, υ Φ, φ Χ, χ Ψ, ψ Ω, ω

kappa lambda mu nu xi omicron epi rho sigma tau upsilon phi chi psi omega

Corrosion.. Fundamentals, Testing, and Protection (10ed., ASM, 2003)(ISBN 0871707055)(2597s) opt

Corrosion and Protection

Corrosion and protection

ASM Handbook: Volume 8: Mechanical Testing and Evaluation (Asm Handbook) (Asm Handbook) (Asm Handbook) (Asm Handbook)

Pipeline Corrosion and Cathodic Protection

Corrosion and Protection (Engineering Materials and Processes)

Fundamentals of Electrochemical Corrosion

Handbook of Cathodic Corrosion Protection

Corrosion Prevention and Protection Practical Solutions

Plastics Fundamentals, Properties, and Testing

Fundamentals of metallic corrosion: atmospheric and media corrosion of metals

Mechanical Wear Fundamentals and Testing

Fundamentals of Corrosion: Mechanisms, Causes, and Preventative Methods (Corrosion technology)

Corrosion and Corrosion Control

Software Testing Fundamentals

ASM Handbook: Volume 7: Powder Metal Technologies and Applications (Asm Handbook) (Asm Handbook) (Asm Handbook)

CORROSION of LININGS and COATINGS: Cathodic and Inhibitor Protection and Corrosion Monitoring (Corrosion Engineering Handbook, Second Edition)

Organic Chemistry Demystified opt

Coatings Technology: Fundamentals, Testing, and Processing Techniques

Electromagnetic Shielding and Corrosion Protection for Aerospace Vehicles

ASM Handbook: Volume 6: Welding, Brazing, and Soldering (Asm Handbook)

Corrosion

Corrosion

Corrosion-Resistant Linings and Coatings (Corrosion Technology)

ASM Handbook: Volume 6: Welding, Brazing, and Soldering (Asm Handbook)

Corrosion and Corrosion Control 4th Edition

Principles of Corrosion Engineering and Corrosion Control

Principles of Corrosion Engineering and Corrosion Control

The Fundamentals of Mixed Signal Testing

Mechanical Wear Fundamentals and Testing, Revised and Expanded

Mechanical Wear Fundamentals and Testing, Revised and Expanded

Corrosion.. Fundamentals, Testing, and Protection (10ed., ASM, 2003)(ISBN 0871707055)(2597s) opt

Corrosion and Protection

Corrosion and protection

ASM Handbook: Volume 8: Mechanical Testing and Evaluation (Asm Handbook) (Asm Handbook) (Asm Handbook) (Asm Handbook)

Pipeline Corrosion and Cathodic Protection

Corrosion and Protection (Engineering Materials and Processes)

Fundamentals of Electrochemical Corrosion

Handbook of Cathodic Corrosion Protection

Corrosion Prevention and Protection Practical Solutions

Plastics Fundamentals, Properties, and Testing

Fundamentals of metallic corrosion: atmospheric and media corrosion of metals

Mechanical Wear Fundamentals and Testing

Fundamentals of Corrosion: Mechanisms, Causes, and Preventative Methods (Corrosion technology)

Corrosion and Corrosion Control

Software Testing Fundamentals

ASM Handbook: Volume 7: Powder Metal Technologies and Applications (Asm Handbook) (Asm Handbook) (Asm Handbook)

CORROSION of LININGS and COATINGS: Cathodic and Inhibitor Protection and Corrosion Monitoring (Corrosion Engineering Handbook, Second Edition)

Organic Chemistry Demystified opt

Coatings Technology: Fundamentals, Testing, and Processing Techniques

Electromagnetic Shielding and Corrosion Protection for Aerospace Vehicles

ASM Handbook: Volume 6: Welding, Brazing, and Soldering (Asm Handbook)

Corrosion

Corrosion

Corrosion-Resistant Linings and Coatings (Corrosion Technology)

ASM Handbook: Volume 6: Welding, Brazing, and Soldering (Asm Handbook)

Corrosion and Corrosion Control 4th Edition

Principles of Corrosion Engineering and Corrosion Control

Principles of Corrosion Engineering and Corrosion Control

The Fundamentals of Mixed Signal Testing

Mechanical Wear Fundamentals and Testing, Revised and Expanded

Mechanical Wear Fundamentals and Testing, Revised and Expanded

Recommend Documents